Background:

Hodgkin lymphoma survivors treated with infradiaphragmatic radiotherapy (IRT) and/or procarbazine have an increased risk of developing colorectal cancer. We investigated the cost-effectiveness of colorectal cancer surveillance in Dutch Hodgkin lymphoma survivors to determine the optimal surveillance strategy for different Hodgkin lymphoma subgroups.

Methods:

The Microsimulation Screening Analysis-Colon model was adjusted to reflect colorectal cancer and other-cause mortality risk in Hodgkin lymphoma survivors. Ninety colorectal cancer surveillance strategies were evaluated varying in starting and stopping age, interval, and modality [colonoscopy, fecal immunochemical test (FIT, OC-Sensor; cutoffs: 10/20/47 μg Hb/g feces), and multi-target stool DNA test (Cologuard)]. Analyses were also stratified per primary treatment (IRT and procarbazine or procarbazine without IRT). Colorectal cancer deaths averted (compared with no surveillance) and incremental cost-effectiveness ratios (ICER) were primary outcomes. The optimal surveillance strategy was identified assuming a willingness-to-pay threshold of €20,000 per life-years gained (LYG).

Results:

Overall, the optimal surveillance strategy was annual FIT (47 μg) from age 45 to 70 years, which might avert 70% of colorectal cancer deaths in Hodgkin lymphoma survivors (compared with no surveillance; ICER:€18,000/LYG). The optimal surveillance strategy in Hodgkin lymphoma survivors treated with procarbazine without IRT was biennial FIT (47 μg) from age 45 to 70 years (colorectal cancer mortality averted 56%; ICER:€15,000/LYG), and when treated with IRT and procarbazine, annual FIT (47 μg) surveillance from age 40 to 70 was most cost-effective (colorectal cancer mortality averted 75%; ICER:€13,000/LYG).

Conclusions:

Colorectal cancer surveillance in Hodgkin lymphoma survivors is cost-effective and should commence earlier than screening occurs in population screening programs. For all subgroups, FIT surveillance was the most cost-effective strategy.

Impact:

Colorectal cancer surveillance should be implemented in Hodgkin lymphoma survivors.

This article is featured in Highlights of This Issue, p. 2093

Hodgkin lymphoma survivors treated with infradiaphragmatic radiotherapy (IRT) and/or procarbazine-containing chemotherapy have a higher risk of developing colorectal cancer in comparison with the general population with a relative risk between 2 and 7 (1–6). Overtime, the treatment for Hodgkin lymphoma changed resulting in a better survival and therefore these patients have a higher chance of developing late adverse events, among which the development of second primary malignancies (7). Colorectal cancer surveillance may be indicated as a higher prevalence of (advanced) adenomas and serrated polyps has been shown (6). Colonoscopy surveillance has the possibility to remove benign precursor lesions and to detect colorectal cancer in an earlier stage. Therefore, colorectal cancer surveillance potentially could decrease colorectal cancer incidence and improve colorectal cancer related mortality in Hodgkin lymphoma survivors. However, tailored recommendations are lacking for this population.

It is unknown whether surveillance could lead to a clinically meaningful reduction in colorectal cancer mortality for Hodgkin lymphoma survivors diagnosed at an adult age. In the United States, colonoscopy surveillance is recommended for young adults who survived childhood cancer (including Hodgkin lymphoma) diagnosed before the age of 21 years and were previously treated with IRT. However, so far patients treated with procarbazine-containing chemotherapy have not been included in the recommendations (8). Moreover, despite surveillance being recommended, participation rates in colonoscopy surveillance have been low among cancer survivors. An alternative colorectal cancer surveillance modality would be a noninvasive stool test, like a fecal immunochemical test (FIT) or multi-target stool DNA test (mt-sDNA), which were found to identify advanced neoplasia among Hodgkin lymphoma survivors diagnosed at an adult age (9). However, it is still unclear from which age to start colorectal cancer surveillance and which surveillance modality—colonoscopy or stool test—would be optimal and if it should vary according to previous Hodgkin lymphoma treatment.

In this study, we performed a cost-effectiveness analysis using a microsimulation modelling approach to determine the optimal colorectal cancer surveillance strategy for Hodgkin lymphoma survivors in the Netherlands, including both colonoscopy and stool test surveillance.

We adjusted well-established Microsimulation Screening Analysis-Colon (MISCAN-Colon; refs. 8, 10–13) model to reflect the Hodgkin lymphoma survivor population. Subsequently the model was used to evaluate benefits, harms, and costs of a range of potential surveillance strategies. An incremental cost-effectiveness analyses were performed to determine which strategy is optimal.

MISCAN-colon model

We adjusted the existing MISCAN-Colon model for the Dutch general population to reflect the colorectal cancer and the other-cause mortality risk observed among Hodgkin lymphoma survivors. MISCAN-Colon is a validated microsimulation model described extensively in previous papers (8, 10–13).

Adaptions of the MISCAN-colon model to Hodgkin lymphoma survivors

The adjustments for the Hodgkin lymphoma population were based on a large Dutch cohort study that aimed to prospectively assess the prevalence of colorectal neoplasia in Hodgkin lymphoma survivors, selecting. Five-year Hodgkin lymphoma survivors with first treatment performed between 1965 and 1995. The treatment strategies of that study were in accordance with treatment protocols of the European Organisation for Research and Treatment of Cancer. However, treatments for recurrence were not standardized. The adjustments are described in Table 1. In our analysis, we adjusted an existing version of the MISCAN-Colon model to reflect colorectal cancer risk and life expectancy of Hodgkin lymphoma survivors (Table 1). We adjusted our model parameters using the standardized incidence ratio (SIR) of colorectal cancer observed in a Dutch cohort of Hodgkin lymphoma survivors. Those rates detected a 3.0-fold increased risk of developing colorectal cancer for Hodgkin lymphoma survivors (regardless of the Hodgkin lymphoma treatment strategy, but treatment included IRT and/or procarbazine-containing chemotherapy) compared with the general population (1). The MISCAN-Colon model specifically simulates the adenoma–carcinoma sequence (14, 15), and does not directly simulate serrated lesions. In the current study, we assumed that the progression times from adenoma onset to colorectal cancer progression among Hodgkin lymphoma survivors were comparable with the general population. However, adenomas were assumed to be more often located in the proximal colon as seen in the cohort of Dutch Hodgkin lymphoma survivors. Our model assumptions (for the natural history of colorectal cancer) were tested replicating observed Dutch and worldwide data on Hodgkin lymphoma survivors (Supplementary Figs. S1 and S2; refs. 6, 16). In this modelling exercise (model validation), we tested both assumptions related to the causes of the higher colorectal cancer risk [as consequence of a higher onset of adenoma (base case analysis) versus as consequence of a combination of higher adenoma onset and faster progression from adenoma to carcinoma (sensitivity analysis)] as described in the Methods of our study. The results of the stool tests (FIT and Mt-sDNA) were based on a prospective study, which evaluated the diagnostic accuracy in Hodgkin lymphoma survivors (9).

Table 1.

Key modelling assumptions.

Input parameterModel assumptionsOne-way sensitivity analyses
Demography 
 All-cause mortality Dutch lifetables (2016; ref. 43), adjusted assuming 5.2-fold increased all-cause mortality in Hodgkin lymphoma survivors 1. 
  Dutch lifetables (2016; ref. 43), adjusted assuming the following increased risks in all-cause mortality according years since Hodgkin lymphoma diagnosis: (25
  10–14 years: RR = 7.2 
  15–19 years: RR = 4.7 
  20–24 years: RR = 4.3 
  25–29 years: RR = 5.0 
  ≥30 years: RR = 6.9 
  2. 
  Dutch lifetables (2016; ref. 43), adjusted assuming 3.12-fold increased all-cause mortality in Hodgkin lymphoma survivors (26
Natural history 
 Adenoma onset Age-dependent (nonhomogeneous Poisson) with more frequent adenoma (assumed after diagnosis of Hodgkin lymphoma, age 25 years) adjusted according to colorectal cancer risks observed in Hodgkin lymphoma survivors:
Entire cohort of Hodgkin lymphoma survivors: RR = 3.4λ;
Hodgkin lymphoma survivors with IRT + Procarbazine: RR = 7.12λ;
Hodgkin lymphoma survivors treated with procarbazine without IRT: RR = 2.1λ
3.
Entire cohort of Hodgkin lymphoma survivors combined: RR = 1.75λ;
Hodgkin lymphoma survivors with IRT + Procarbazine: RR = 3.65λ;
Hodgkin lymphoma survivors treated with procarbazine without IRT: RR = 1.1λ.
Assuming a shorter adenoma state duration compared with the general population: Exp(λ = 70)λ
4.
According to Rigter and colleagues (2019), Supplementary Fig. S2:
Entire cohort of Hodgkin lymphoma survivors: RR = 4.85λ;
Hodgkin lymphoma survivors with IRT + Procarbazine: RR = 7.16λ;
Hodgkin lymphoma survivors treated with procarbazine without IRT: RR = 3.1λ 
 Adenoma localization Rectum: 7.9%; Sigmoid: 11.45%; Descending:10.75%; Transverse: 31.85%; Ascending: 26.05%; and Cecum:12% (65. 
  Rectum: 26.38%; Rectosigmoid: 9.12%; Sigmoid: 26.37%; Descending:6%; Transverse: 9.01%; Ascending: 8.85%; and Cecum:14.27% (29
 Adenoma progression 
  State transitions Age-dependent  
  State durations, years (total) Exp(λ = 140)λ See 3. 
 Cancer progression (preclinical)   
  Stage transitions Age-dependent  
  Stage durations, years Exp(λ = 2.5)  
 Colorectal cancer survival Age-/Stage-/Localization-dependent 6. 
  1.33-fold lower compared with Dutch general population with a colorectal cancer diagnosis (27
FIT and sMT-DNA performance   
 FIT   
 Sensitivity***, % 10 μg Hb/g feces 20μg Hb/g feces 47 μg Hb/g feces MT-sDNA 7.
Systematic FIT negative results were assumed (28
  adenomas <10mm 
8.
Sensitivity for adenomas (6–9mm, %; ref. 29):
10 μg Hb/g feces = 9.6;
20 μg Hb/g feces = 4.4;
47μg Hb/g feces = 2.5. 
  adenomas ≥10mm 26.5 18.5 12.6 31.1  
  malignant neoplasia (early) §§29 65 52.5 50 97  
  malignant neoplasia (late) §§29 90 83.5 82.5 86  
 Specificity, % 91 95 96 62  
Colonoscopy performance 
 Sensitivity, % 
  adenomas 0–5mm 75  
  adenomas 6–9mm 85  
  adenomas ≥10mm 95  
  malignant neoplasia 95  
 Specificity, % 86  
 Complete colonoscopy examination, % 100 (69. 
  92 (29
 Complication rates, % with polypectomy§ Age-dependent  
  Fatal complications|| 0.000329  
 without polypectomy  
Costs,   
 FIT 15  
 sMT-DNA 604*  
 Colonoscopy   
  with polypectomy 887  
  without polypectomy 679  
 Complications#,**,†† 3,488  
 Per life-year with cancer care   
  Initial year, stage I–IV 15,222–30,444 10–11. 50% higher and 100% higher 
  Ongoing, stage I–IV 414  
  Terminal year (colorectal cancer death), stage I-IV 21,311–30,444  
  Terminal year (other causes), stage I–IV 5,358–17,049  
Discounting rates (Cost-effectiveness analysis)  12. 
 Benefits 3% 1.5% 
 Costs 3% 4% 
Input parameterModel assumptionsOne-way sensitivity analyses
Demography 
 All-cause mortality Dutch lifetables (2016; ref. 43), adjusted assuming 5.2-fold increased all-cause mortality in Hodgkin lymphoma survivors 1. 
  Dutch lifetables (2016; ref. 43), adjusted assuming the following increased risks in all-cause mortality according years since Hodgkin lymphoma diagnosis: (25
  10–14 years: RR = 7.2 
  15–19 years: RR = 4.7 
  20–24 years: RR = 4.3 
  25–29 years: RR = 5.0 
  ≥30 years: RR = 6.9 
  2. 
  Dutch lifetables (2016; ref. 43), adjusted assuming 3.12-fold increased all-cause mortality in Hodgkin lymphoma survivors (26
Natural history 
 Adenoma onset Age-dependent (nonhomogeneous Poisson) with more frequent adenoma (assumed after diagnosis of Hodgkin lymphoma, age 25 years) adjusted according to colorectal cancer risks observed in Hodgkin lymphoma survivors:
Entire cohort of Hodgkin lymphoma survivors: RR = 3.4λ;
Hodgkin lymphoma survivors with IRT + Procarbazine: RR = 7.12λ;
Hodgkin lymphoma survivors treated with procarbazine without IRT: RR = 2.1λ
3.
Entire cohort of Hodgkin lymphoma survivors combined: RR = 1.75λ;
Hodgkin lymphoma survivors with IRT + Procarbazine: RR = 3.65λ;
Hodgkin lymphoma survivors treated with procarbazine without IRT: RR = 1.1λ.
Assuming a shorter adenoma state duration compared with the general population: Exp(λ = 70)λ
4.
According to Rigter and colleagues (2019), Supplementary Fig. S2:
Entire cohort of Hodgkin lymphoma survivors: RR = 4.85λ;
Hodgkin lymphoma survivors with IRT + Procarbazine: RR = 7.16λ;
Hodgkin lymphoma survivors treated with procarbazine without IRT: RR = 3.1λ 
 Adenoma localization Rectum: 7.9%; Sigmoid: 11.45%; Descending:10.75%; Transverse: 31.85%; Ascending: 26.05%; and Cecum:12% (65. 
  Rectum: 26.38%; Rectosigmoid: 9.12%; Sigmoid: 26.37%; Descending:6%; Transverse: 9.01%; Ascending: 8.85%; and Cecum:14.27% (29
 Adenoma progression 
  State transitions Age-dependent  
  State durations, years (total) Exp(λ = 140)λ See 3. 
 Cancer progression (preclinical)   
  Stage transitions Age-dependent  
  Stage durations, years Exp(λ = 2.5)  
 Colorectal cancer survival Age-/Stage-/Localization-dependent 6. 
  1.33-fold lower compared with Dutch general population with a colorectal cancer diagnosis (27
FIT and sMT-DNA performance   
 FIT   
 Sensitivity***, % 10 μg Hb/g feces 20μg Hb/g feces 47 μg Hb/g feces MT-sDNA 7.
Systematic FIT negative results were assumed (28
  adenomas <10mm 
8.
Sensitivity for adenomas (6–9mm, %; ref. 29):
10 μg Hb/g feces = 9.6;
20 μg Hb/g feces = 4.4;
47μg Hb/g feces = 2.5. 
  adenomas ≥10mm 26.5 18.5 12.6 31.1  
  malignant neoplasia (early) §§29 65 52.5 50 97  
  malignant neoplasia (late) §§29 90 83.5 82.5 86  
 Specificity, % 91 95 96 62  
Colonoscopy performance 
 Sensitivity, % 
  adenomas 0–5mm 75  
  adenomas 6–9mm 85  
  adenomas ≥10mm 95  
  malignant neoplasia 95  
 Specificity, % 86  
 Complete colonoscopy examination, % 100 (69. 
  92 (29
 Complication rates, % with polypectomy§ Age-dependent  
  Fatal complications|| 0.000329  
 without polypectomy  
Costs,   
 FIT 15  
 sMT-DNA 604*  
 Colonoscopy   
  with polypectomy 887  
  without polypectomy 679  
 Complications#,**,†† 3,488  
 Per life-year with cancer care   
  Initial year, stage I–IV 15,222–30,444 10–11. 50% higher and 100% higher 
  Ongoing, stage I–IV 414  
  Terminal year (colorectal cancer death), stage I-IV 21,311–30,444  
  Terminal year (other causes), stage I–IV 5,358–17,049  
Discounting rates (Cost-effectiveness analysis)  12. 
 Benefits 3% 1.5% 
 Costs 3% 4% 

Abbreviation: RR, relative risk.

λ The combination of increased adenoma onset and short adenoma state duration resulted in a risk of colorectal cancer (compared with Dutch general population), respectively, of 3.0-fold higher in the entire cohort of Hodgkin lymphoma survivors, 2.0-fold in Hodgkin lymphoma survivors treated with procarbazine without IRT, and 5.7-fold in Hodgkin lymphoma survivors treated with IRT and procarbazine (Fig. 1; ref. 6).

The sensitivity of colonoscopy for the detection of adenomas and colorectal cancer within the reach of the endoscope was obtained from a systematic review on miss rates seen in tandem colonoscopy studies (44).

Specificity for colonoscopy is therefore based on an adenoma prevalence study of patients undergoing surveillance colonoscopy (45).

§Age-specific risks for complications of colonoscopy requiring a hospital admission or emergency department visit were obtained from a study by Warren and colleagues (46).

||The mortality rate associated with colonoscopies with a polypectomy was derived by multiplying the risk for a perforation obtained from a study by Warren and colleagues (46) by the risk for death given a perforation obtained from a study by Gatto and colleagues (47).

Costs are presented in Euro.

#Serious gastrointestinal (GI) complications included perforations, gastrointestinal bleeding, or transfusions.

**Other gastrointestinal (GI) complications included paralytic ileus, nausea and vomiting, dehydration, or abdominal pain.

††Cardiovascular complications included myocardial infarction or angina, arrhythmias, congestive heart failure, cardiac or respiratory arrest, syncope, hypotension, or shock.

§§FIT sensitivity for malignant neoplasia were informed using the study of Goede SL and colleagues (2013). For FIT 47 μg Hb/g feces those sensitivity values were assumed equal to those provided for FIT 40 μg Hb/g feces. (29).

*maximum reimbursement cost in US as assumed in Lew and colleagues (2018) IJC (20).

***Sensitivities were per-lesion.

Adjustments to reflect the Hodgkin lymphoma population: 3 cohorts

Briefly, we used the SIR of colorectal cancer observed in a Dutch cohort of 5-year Hodgkin lymphoma survivors to assume a 3.0-fold increased risk of colorectal cancer in Hodgkin lymphoma survivors (for the entire cohort including all Hodgkin lymphoma treatment strategies including IRT and/or procarbazine-containing chemotherapy) compared with the general population (Fig. 1; ref. 1). We did not assume changes in risk over calendar time. In our model, we assumed that the higher colorectal cancer risk was a consequence of a higher incidence of adenomas. We assumed the same adenoma incidence as the Dutch general population before Hodgkin lymphoma diagnosis and treatment (from age 0–24 years), and increased adenoma incidence after that. Model validations are reported in Supplementary Document (Supplementary Figs. S1 and S2). We also adjusted the model to consider the 5.2 times higher risk of death for all causes (excluding anal and colorectal cancer mortality) observed among Hodgkin lymphoma survivors (compared with the general population in same age, gender, and calendar period; ref. 7).

Figure 1.

Simulated and expected adenoma prevalence and relative risks for colorectal cancer (compared with average risk individuals) among Hodgkin lymphoma (HL) survivors. This figure represents the validation of the data of HL survivors against published data for the entire cohort (A), HL survivors treated with procarbazine without IRT (B) and HL survivors treated with IRT and procarbazine (C). Simulated outcomes were computed assuming no surveillance.

Figure 1.

Simulated and expected adenoma prevalence and relative risks for colorectal cancer (compared with average risk individuals) among Hodgkin lymphoma (HL) survivors. This figure represents the validation of the data of HL survivors against published data for the entire cohort (A), HL survivors treated with procarbazine without IRT (B) and HL survivors treated with IRT and procarbazine (C). Simulated outcomes were computed assuming no surveillance.

Close modal

The different treatment strategies for Hodgkin lymphoma resulted in different SIRs for developing colorectal cancer (1). Compared with the general population, Hodgkin lymphoma survivors treated with procarbazine without IRT had a 2.0-fold higher SIR for colorectal cancer, whereas in those treated with IRT and procarbazine the risk was 5.7-fold higher (1). We, therefore, also performed separate analyses considering differences in colorectal cancer risk based on Hodgkin lymphoma treatment (increasing or decreasing the risk accordingly). In those analyses, we assumed no difference in all-cause mortality by Hodgkin lymphoma treatment (i.e., 5.2-fold higher than the general population; ref. 7). Validation of these two separate model versions was performed and reported in Fig. 1.

Surveillance strategies simulated

We performed a cost-effectiveness analysis for each Hodgkin lymphoma survivor group to explore possible reasonable options to determine the most optimal surveillance strategy. We evaluated benefits and costs of 90 different surveillance strategies (including no surveillance) varying in test modality (colonoscopy, FIT with different positivity cutoffs, or mt-sDNA), age to start (35, 40, 45, 50 years), intervals (3, 5, and 10 years for colonoscopy, and 1 and 2 years for stool tests), and age to end (70 or 75 years). These variations were evaluated to determine the most beneficial strategy for the different subgroups. Test characteristics of the stool tests for detecting advanced neoplasia were based on diagnostic analysis previously performed in Hodgkin lymphoma survivors treated with IRT and/or procarbazine-containing chemotherapy who prospectively underwent a colonoscopy and performed stool tests prior to colonoscopy (Table 1; ref. 9). For FIT (OC-Sensor, Eiken Chemical, Tokyo, Japan), three different predetermined cutoffs were evaluated, specifically 10, 20, 47 μg Hb/g feces. The positivity at the mt-sDNA test (Cologuard, Exact Sciences Corporation, Madison, United States) was classified as described in previous studies (9, 17). Participants with a positive stool test were simulated to undergo a colonoscopy (6). We assumed that the completion rate of colonoscopy was 100% and the complication rate was similar to the general population. We assumed 100% participation in all surveillance and diagnostic follow-up.

Costs

We applied a modified societal perspective for the cost-effectiveness analysis, including patient time costs but no other indirect costs (i.e., traveling). Cost for colonoscopy and FIT surveillance has been informed from the monitoring report of the Dutch FIT organized screening programme (18). As information on the cost for the mt-sDNA test are lacking in the Netherlands, we assumed the maximum out-of-pocket cost ($649, 2017) of Cologuard in U.S. market (19, 20). Costs for treatment and care of colorectal cancer have previously been published (21). All costs were updated to the year 2019 using the cost price index from the Dutch Health Care Authority (22).

Outcomes

We simulated three cohorts of 10 million Hodgkin lymphoma survivors aged 35 years old in 2019 (with Hodgkin lymphoma diagnosed at age of 25 years) for the three treatment categories (entire cohort of Hodgkin lymphoma survivors including all Hodgkin lymphoma treatment strategies, Hodgkin lymphoma survivors treated with procarbazine without IRT, and Hodgkin lymphoma survivors treated with IRT and procarbazine). We simulated three cohorts of individuals all born in the same year. Although the actual number of Hodgkin lymphoma survivors in the Netherlands evidently is not in that order of magnitude (8, 11), the large cohort sample size was chosen to guarantee stable model outcomes in our simulations. To endorse generalizability to Hodgkin lymphoma survivor populations of different sizes all outcomes are reported per 1.000 survivors aged 35 years in 2019. Age 35 was chosen because the simulated increase in adenoma incidence from age 25 years onwards would require at least 10 years for these adenomas (caused by Hodgkin lymphoma treatment) to result in an increase in colorectal cancer incidence. As the information used to inform the model was limited, the increase in the adenoma incidence was assumed to not change according to period of Hodgkin lymphoma diagnosis.

For each surveillance strategy, the surveillance effectiveness [i.e., number of colorectal cancer deaths prevented, relative colorectal cancer mortality reduction and life-years gained (LYG)] and resources (colonoscopies, FIT, mt-sDNA test and cost) were analyzed, discounting the LYG and cost at the conventional 3% annual discount rate (Supplementary Tables S1–S3). We calculated the number of colonoscopies needed to prevent a colorectal cancer death by dividing the total number of colonoscopies performed (per 1,000) by the number of colorectal cancer deaths prevented per 1,000 Hodgkin lymphoma survivors screened, referred to as number needed to screen (NNS, Supplementary Tables S4–S6). For each group of Hodgkin lymphoma survivors, we also predicted colorectal cancer deaths and total costs simulating two existing surveillance recommendations indicated in the Netherlands: (i) the screening strategy for the Dutch general population (biennial FIT, i.e., once every 2 years, from age 55 to 75 years, 47 μg Hb/g feces); and (ii) the surveillance strategy for individuals with a family history of colorectal cancer (colonoscopy surveillance repeated every 5 years from age 45–75 years).

Cost-effectiveness analyses

The optimal colorectal cancer surveillance strategy in Hodgkin lymphoma survivors was determined by first excluding strategies that were more expensive and less effective than (combinations of) other simulated strategies (23). For the remaining strategies (defined as ‘efficient strategies’), we calculated the incremental cost-effectiveness ratio (ICER) by comparing the ratio between additional costs and LYG to the next less expensive efficient strategy. The optimal strategy was defined as the most effective strategy with an ICER below the willingness-to-pay threshold of €20,000 per LYG (21, 24). Strategies with an ICER exceeding €20,000 were considered not cost-effective. A separate analysis was performed excluding the stool tests to evaluate which colonoscopy surveillance program was most cost-effective (results reported and discussed in the Supplementary Methods and Supplementary Tables S7–S9).

Sensitivity analyses

Multiple one-way sensitivity analyses were performed to reinforce the results under a variety of assumptions (Table 1). Those assumptions included an adjustment in the lifetables including different relative risks for different intervals after Hodgkin lymphoma treatment (25); another adjustment in the lifetables based on Anderson et al (26); higher colorectal cancer risk as shown in a prospective study in which Hodgkin lymphoma survivors underwent a colonoscopy and a higher prevalence of advanced neoplasia was detected; Supplementary Fig. S2 (6); different colorectal cancer localization (in line with the general Dutch population; ref. 6); a 1.33 lower colorectal cancer relative survival (27); systematic FIT negative results (28); FIT sensitivity for medium adenomas (6–9 mm) assumed as reported for the Dutch general population (29); a different assumption for the pathway to higher colorectal cancer risk in Hodgkin lymphoma survivors (colorectal cancer risk caused by a higher adenoma onset in combination with a twice-faster adenoma progression; Fig. 1); a lower complete colonoscopy examination rate (92% instead of 100%); higher costs for colorectal cancer treatment and care (50% and 100% higher); and 4% discount rate for costs and 1.5% for benefits as recommended by the Dutch Ministry of Health (30).

Data availability

The data generated in this study are available within the article and its Supplementary Data files. Detailed data generated in this study about the MISCAN model is available upon request from the corresponding author.

In the entire Hodgkin lymphoma survivor cohort (not stratified by Hodgkin lymphoma treatment), 26 colorectal cancer deaths per 1000 Hodgkin lymphoma survivors (starting aged 35 years in 2019) were predicted over a lifetime in the absence of surveillance (Table 2). Up to 49% of those colorectal cancer deaths may be averted with the recommended screening strategy for the Dutch general population (Fig. 2) with biennial FIT 47 μg Hb/g feces between 55 to 75 years of age at the total costs of € 1.1 million per 1,000 Hodgkin lymphoma survivors (NNS = 75, data not included). The surveillance strategy indicated for individuals with family history of colorectal cancer being primary colonoscopy surveillance from age 45 years can prevent up to 80% of colorectal cancer mortality in Hodgkin lymphoma survivors, however, at higher costs (total costs €2.4 million per 1,000 Hodgkin lymphoma survivors, NNS = 222). The most optimal cost-effective colorectal cancer surveillance strategy was annual FIT surveillance from age 45 to 70 years using a positivity cut-off threshold of 47 μg Hb/g feces, which prevented up to 70% of colorectal cancer mortality, however, at lower costs than the previous colonoscopy strategy (compared with no surveillance; total costs € 1.4 million per 1,000 Hodgkin lymphoma survivors; NNS = 75; ICER = €18,000 per LYG; Fig. 3, Table 2, and Supplementary Table S4 and S10).

Table 2:

Efficient surveillance strategies among Hodgkin lymphoma survivors per primary cancer treatment (base case analysis).

Outcomes per 1,000 Hodgkin lymphoma survivors free of colorectal cancer diagnosis and aged 35 years in 2019 (3%)*Reductions:
Surveillance strategiesFITsScr. COLsDiag. COLsSurv. COLsTotal COLsCompl.Colorectal cancersColorectal cancer deathsColorectal cancer careLYGTotal costsNet costsIncidence (%)Mortality (%)ICER (*1,000)
Entire cohort of Hodgkin lymphoma survivors 
No Surveillance 33 33 73 26 214 966 
FIT47, 50–70, 2 years 3,457 199 145 344 63 12 274 38 1,161 196 13 54 
FIT47, 45–70, 2 years 4,936 254 198 452 60 11 273 46 1,224 258 18 59 
FIT20, 45–70, 2 years 4,779 342 258 599 54 10 253 49 1,271 306 26 63 13 
FIT47, 45–70, 1 years 8,957 0 408 293 701 3 51 8 254 55 1,373 407 30 70 18 
FIT47, 40–70, 1 years 12,490 536 361 897 49 246 61 1,537 571 33 72 29 
FIT20, 40–70, 1 years 11,975 740 447 1,187 43 220 64 1,681 716 41 75 53 
FIT20, 35–70, 1 years 16,252 976 510 1,486 42 214 67 1,943 978 42 76 75 
FIT20, 35–75, 1 years 16,611 997 511 1,508 43 215 67 1,967 1,001 42 78 87 
FIT10, 35–70, 1 years 15,462 1,491 634 2,125 37 188 70 2,339 1,374 49 79 130 
FIT10, 35–75, 1 years 15,809 1,523 634 2,158 37 189 70 2,370 1,405 49 80 148 
COL, 35–70, 3 years 4,864 1,956 6,822 22 112 80 5,337 4,371 70 88 308 
COL, 35–75, 3 years 5,003 1,956 6,960 21 112 80 5,437 4,471 71 89 439 
Hodgkin lymphoma survivors treated with procarbazine without IRT 
No Surveillance 22 22 49 17 141 637 
FIT47, 50–70, 2 years 3,568 183 108 291 44 184 23 847 210 11 52 
FIT47, 45–70, 2 years 5,075 0 240 147 387 2 42 8 183 28 916 279 15 56 15 
FIT20, 45–70, 2 years 4,945 331 194 525 38 170 30 977 339 23 61 25 
FIT47, 45–70, 1 years 9,340 402 222 623 36 171 35 1,091 453 27 68 28 
FIT47, 40–70, 1 years 12,963 535 273 808 34 167 38 1,257 620 30 70 45 
FIT47, 40–75, 1 years 13,362 550 274 825 35 169 39 1,279 642 29 72 71 
FIT20, 40–70, 1 years 12,528 753 343 1,096 30 149 40 1,418 781 39 74 88 
FIT47, 35–75, 1 years 17,784 713 314 1,027 34 166 41 1,479 842 31 74 91 
FIT20, 35–75, 1 years 17,273 1,018 393 1,411 29 146 43 1,705 1,068 40 77 115 
FIT10, 35–70, 1 years 16,223 1,550 494 2,043 25 127 45 2,103 1,465 49 79 182 
FIT10, 35–75, 1 years 16,597 1,585 494 2,079 25 127 45 2,135 1,498 49 80 191 
COL, 35–70, 3 years 5,209 1,611 6,822 14 73 52 5,176 4,539 72 88 461 
COL, 35–75, 3 years 5,359 1,611 6,972 14 73 52 5,284 4,647 72 90 623 
Hodgkin lymphoma survivors treated with IRT and procarbazine chemotherapy 
No Surveillance 59 59 127 47 392 1,753 
FIT20, 50–70, 2 years 3,094 291 282 573 96 18 463 81 1,867 113 24 61 
FIT20, 45–70, 2 years 4,465 367 387 754 90 16 451 100 1,929 175 29 67 
FIT10, 45–70, 2 years 4,197 501 494 995 80 13 419 108 1,998 244 37 72 
FIT47, 45–70, 1 years 8,255 426 434 860 85 13 451 109 2,006 253 33 72 10 
FIT47, 40–70, 1 years 11,623 0 545 540 1,086 5 81 12 435 121 2,156 403 36 75 13 
FIT20, 40–70, 1 years 10,997 723 655 1,378 72 11 391 125 2,268 515 43 77 25 
FIT20, 35–70, 1 years 15,099 945 753 1,698 70 10 377 132 2,528 774 45 78 41 
FIT20, 35–75, 1 years 15,414 963 754 1,717 70 10 378 132 2,549 796 44 79 69 
FIT10, 35–70, 1 years 14,131 1,393 916 2,309 63 336 137 2,877 1,123 50 80 73 
FIT10, 35–75, 1 years 14,436 1,422 916 2,338 63 337 137 2,904 1,151 50 81 115 
COL, 35–70, 3 years 4,289 2,516 6,810 11 41 217 153 5,700 3,947 68 88 173 
COL, 35–75, 3 years 4,410 2,516 6,930 12 40 218 153 5,787 4,034 68 89 283 
Outcomes per 1,000 Hodgkin lymphoma survivors free of colorectal cancer diagnosis and aged 35 years in 2019 (3%)*Reductions:
Surveillance strategiesFITsScr. COLsDiag. COLsSurv. COLsTotal COLsCompl.Colorectal cancersColorectal cancer deathsColorectal cancer careLYGTotal costsNet costsIncidence (%)Mortality (%)ICER (*1,000)
Entire cohort of Hodgkin lymphoma survivors 
No Surveillance 33 33 73 26 214 966 
FIT47, 50–70, 2 years 3,457 199 145 344 63 12 274 38 1,161 196 13 54 
FIT47, 45–70, 2 years 4,936 254 198 452 60 11 273 46 1,224 258 18 59 
FIT20, 45–70, 2 years 4,779 342 258 599 54 10 253 49 1,271 306 26 63 13 
FIT47, 45–70, 1 years 8,957 0 408 293 701 3 51 8 254 55 1,373 407 30 70 18 
FIT47, 40–70, 1 years 12,490 536 361 897 49 246 61 1,537 571 33 72 29 
FIT20, 40–70, 1 years 11,975 740 447 1,187 43 220 64 1,681 716 41 75 53 
FIT20, 35–70, 1 years 16,252 976 510 1,486 42 214 67 1,943 978 42 76 75 
FIT20, 35–75, 1 years 16,611 997 511 1,508 43 215 67 1,967 1,001 42 78 87 
FIT10, 35–70, 1 years 15,462 1,491 634 2,125 37 188 70 2,339 1,374 49 79 130 
FIT10, 35–75, 1 years 15,809 1,523 634 2,158 37 189 70 2,370 1,405 49 80 148 
COL, 35–70, 3 years 4,864 1,956 6,822 22 112 80 5,337 4,371 70 88 308 
COL, 35–75, 3 years 5,003 1,956 6,960 21 112 80 5,437 4,471 71 89 439 
Hodgkin lymphoma survivors treated with procarbazine without IRT 
No Surveillance 22 22 49 17 141 637 
FIT47, 50–70, 2 years 3,568 183 108 291 44 184 23 847 210 11 52 
FIT47, 45–70, 2 years 5,075 0 240 147 387 2 42 8 183 28 916 279 15 56 15 
FIT20, 45–70, 2 years 4,945 331 194 525 38 170 30 977 339 23 61 25 
FIT47, 45–70, 1 years 9,340 402 222 623 36 171 35 1,091 453 27 68 28 
FIT47, 40–70, 1 years 12,963 535 273 808 34 167 38 1,257 620 30 70 45 
FIT47, 40–75, 1 years 13,362 550 274 825 35 169 39 1,279 642 29 72 71 
FIT20, 40–70, 1 years 12,528 753 343 1,096 30 149 40 1,418 781 39 74 88 
FIT47, 35–75, 1 years 17,784 713 314 1,027 34 166 41 1,479 842 31 74 91 
FIT20, 35–75, 1 years 17,273 1,018 393 1,411 29 146 43 1,705 1,068 40 77 115 
FIT10, 35–70, 1 years 16,223 1,550 494 2,043 25 127 45 2,103 1,465 49 79 182 
FIT10, 35–75, 1 years 16,597 1,585 494 2,079 25 127 45 2,135 1,498 49 80 191 
COL, 35–70, 3 years 5,209 1,611 6,822 14 73 52 5,176 4,539 72 88 461 
COL, 35–75, 3 years 5,359 1,611 6,972 14 73 52 5,284 4,647 72 90 623 
Hodgkin lymphoma survivors treated with IRT and procarbazine chemotherapy 
No Surveillance 59 59 127 47 392 1,753 
FIT20, 50–70, 2 years 3,094 291 282 573 96 18 463 81 1,867 113 24 61 
FIT20, 45–70, 2 years 4,465 367 387 754 90 16 451 100 1,929 175 29 67 
FIT10, 45–70, 2 years 4,197 501 494 995 80 13 419 108 1,998 244 37 72 
FIT47, 45–70, 1 years 8,255 426 434 860 85 13 451 109 2,006 253 33 72 10 
FIT47, 40–70, 1 years 11,623 0 545 540 1,086 5 81 12 435 121 2,156 403 36 75 13 
FIT20, 40–70, 1 years 10,997 723 655 1,378 72 11 391 125 2,268 515 43 77 25 
FIT20, 35–70, 1 years 15,099 945 753 1,698 70 10 377 132 2,528 774 45 78 41 
FIT20, 35–75, 1 years 15,414 963 754 1,717 70 10 378 132 2,549 796 44 79 69 
FIT10, 35–70, 1 years 14,131 1,393 916 2,309 63 336 137 2,877 1,123 50 80 73 
FIT10, 35–75, 1 years 14,436 1,422 916 2,338 63 337 137 2,904 1,151 50 81 115 
COL, 35–70, 3 years 4,289 2,516 6,810 11 41 217 153 5,700 3,947 68 88 173 
COL, 35–75, 3 years 4,410 2,516 6,930 12 40 218 153 5,787 4,034 68 89 283 

Abbreviation: COLs = colonoscopies.

Colorectal cancer cases and colorectal cancer death were not discounted.

Compared with no surveillance.

*Full participation in surveillance and post-colonoscopy surveillance was assumed. Optimal surveillance strategies were reported in bold.

Figure 2.

Colorectal cancer (CRC) deaths and total costs per 1,000 Hodgkin lymphoma (HL) survivors aged 35 years old in 2019 under different surveillance scenarios. The CRC deaths and the total costs per 1,000 HL survivors are shown for the three subgroups of HL survivors (A and B, respectively). The figures display the different optimal surveillance strategies for the entire cohort, HL survivors treated with procarbazine without IRT and HL survivors treated with IRT and procarbazine. The different colors represent the most optimal surveillance strategies for the different HL subgroups as determined in Table 2. Optimal colonoscopy surveillance strategies were determined in Supplementary Tables 7–10. A willingness-to-pay threshold of €20,000 per LYG was assumed in determining the optimal surveillance strategy in each group. COL, colonoscopy.

Figure 2.

Colorectal cancer (CRC) deaths and total costs per 1,000 Hodgkin lymphoma (HL) survivors aged 35 years old in 2019 under different surveillance scenarios. The CRC deaths and the total costs per 1,000 HL survivors are shown for the three subgroups of HL survivors (A and B, respectively). The figures display the different optimal surveillance strategies for the entire cohort, HL survivors treated with procarbazine without IRT and HL survivors treated with IRT and procarbazine. The different colors represent the most optimal surveillance strategies for the different HL subgroups as determined in Table 2. Optimal colonoscopy surveillance strategies were determined in Supplementary Tables 7–10. A willingness-to-pay threshold of €20,000 per LYG was assumed in determining the optimal surveillance strategy in each group. COL, colonoscopy.

Close modal
Figure 3.

Efficient frontier with efficient surveillance strategies for Hodgkin lymphoma (HL) survivors. In this efficient frontier displays the LYG from CRC surveillance against the total costs per 1,000 euro. Total costs and LYG from surveillance were discounted (3% discounting rate) and 100% adherence was assumed for surveillance and diagnostic test. The optimal surveillance strategies are labelled an indicated by arrows. The efficient frontier is shown separately for the entire cohort (A), HL survivors treated with procarbazine chemotherapy without IRT (B), and HL survivors treated with IRT and procarbazine chemotherapy (C). CRC, colorectal cancer.

Figure 3.

Efficient frontier with efficient surveillance strategies for Hodgkin lymphoma (HL) survivors. In this efficient frontier displays the LYG from CRC surveillance against the total costs per 1,000 euro. Total costs and LYG from surveillance were discounted (3% discounting rate) and 100% adherence was assumed for surveillance and diagnostic test. The optimal surveillance strategies are labelled an indicated by arrows. The efficient frontier is shown separately for the entire cohort (A), HL survivors treated with procarbazine chemotherapy without IRT (B), and HL survivors treated with IRT and procarbazine chemotherapy (C). CRC, colorectal cancer.

Close modal

For Hodgkin lymphoma survivors treated with procarbazine without IRT, the model predicted 17 colorectal cancer deaths per 1,000 Hodgkin lymphoma survivors without surveillance. Up to 47% of those deaths could be prevented with the colorectal cancer screening strategy adopted for the Dutch general population (biennial FIT at cutoff 47 55 from age 75 years, total costs €0.8 million per 1,000 Hodgkin lymphoma survivors, NNS = 60, data not included); whereas 80% could be avoided by primary colonoscopy surveillance with a starting age of 45 years, however, at higher costs (total costs = €2.2 million per 1,000 Hodgkin lymphoma survivors, NNS = 336; Fig. 2). However, we found that biennial FIT surveillance (47 μg Hb/g feces) from age 45 and 70 years was the optimal strategy, preventing 56% of colorectal cancer mortality (compared with no surveillance) at an acceptable cost of €15,000/LYG (total costs = € 0.9 million per 1,000 Hodgkin lymphoma survivors; NNS = 79; Fig. 3, Table 2, and Supplementary Table S5 and S11).

In Hodgkin lymphoma survivors treated with IRT and procarbazine, 47 colorectal cancer deaths (per 1,000 Hodgkin lymphoma) were predicted without surveillance. Screening as suggested for the Dutch general population may prevent up to 50% of those deaths (at the costs €1.9 million per 1,000 Hodgkin lymphoma survivors, NNS = 32, data not included); whereas surveillance recommended for individuals with family history of colorectal cancer prevented up to 81% of colorectal cancer (at the cost of €2.8 million per 1,000 Hodgkin lymphoma survivors, NNS = 124; Fig. 2). Nevertheless, annual FIT surveillance (47 μg Hb/g feces) from age 40 and 70 years was optimal, averting 75% of colorectal cancer mortality (compared with no surveillance; total costs = € 2.2 million per 1,000 Hodgkin lymphoma survivors; ICER = €13,000 per LYG; NNS = 56; Fig. 3, Table 2, and Supplementary Table S6 and S12).

For each group of Hodgkin lymphoma survivors, colonoscopy surveillance was estimated not to be cost-effective in comparison with FIT (Supplementary Table S7–S9). Separate analyses were performed excluding stool test surveillance and only including colonoscopy surveillance. This analysis is described in Supplementary Methods and Supplementary Tables S7–S9.

Sensitivity analysis

In all sensitivity analyses, FIT surveillance was most cost-effective for all Hodgkin lymphoma treatment strategies. The optimal cutoff for FIT was quite sensitive to model assumptions. The optimal cutoff changed to a lower cutoff in 31% of the sensitivity analyses. This included among others the analyses which assumed: (i) higher all-cause mortality in the entire Hodgkin lymphoma cohort, (ii) the same FIT sensitivity for medium adenomas as in the general population, and (iii) higher colorectal cancer treatment costs (Table 3). The age range was quite robust, and only changed in the two sensitivity analyses that assumed: (i) a higher colorectal cancer risk in Hodgkin lymphoma survivors, and (ii) the Dutch discounting factors for the entire cohort. The most cost-effective interval changed in a few sensitivity analyses (Table 3).

Table 3.

Optimal surveillance strategies* under different sensitivity analyses.

Hodgkin lymphoma (Entire cohort)Hodgkin lymphoma (Procarbazine without IRT)Hodgkin lymphoma (IRT + Procarbazine)
Basecase analysis FIT47, 45–70, 1 year FIT47, 45–70, 2 years FIT47, 40–70, 1 year 
Sensitivity analyses: 
1. Adjustment lifetables including different relative risks for different intervals after Hodgkin lymphoma treatment FIT20, 45–70, 2 years Unchanged Unchanged 
2. Adjustment in the lifetables based on other Anderson et al Unchanged FIT47, 45–70, 1 year FIT20, 40–70, 1 year 
3. Higher colorectal cancer risk caused by higher adenoma onset FIT47, 40–70, 1 year FIT47, 45–70, 1 year Unchanged 
4. Different adenoma localization Unchanged Unchanged Unchanged 
5. 1.33-fold higher colorectal cancer relative survival Unchanged Unchanged Unchanged 
6. Systematic FIT negative results Unchanged Unchanged Unchanged 
7. FIT sensitivity for medium adenomas (6–9 mm) as general population Unchanged Unchanged FIT10, 40–70, 2 years 
8. Colorectal cancer risk caused directly by a combination of higher adenoma onset and faster adenoma progression Unchanged Unchanged FIT20, 40–70, 1 year 
9. Lower complete colonoscopy examination rate (92%) Unchanged Unchanged Unchanged 
10. Higher colorectal cancer treatment costs (+50%) FIT10, 45–70, 2 years FIT20, 45–70, 2 years FIT10, 40–70, 2 years 
11. Higher colorectal cancer treatment costs (+100%) FIT10, 45–70, 2 years FIT10, 45–70, 2 years FIT20, 40–70, 1 year 
12. Discounting factor (4% costs, 1.5% benefits) FIT47, 40–70, 1 year FIT47, 45–70, 1 year FIT20, 40–70, 1 year 
Hodgkin lymphoma (Entire cohort)Hodgkin lymphoma (Procarbazine without IRT)Hodgkin lymphoma (IRT + Procarbazine)
Basecase analysis FIT47, 45–70, 1 year FIT47, 45–70, 2 years FIT47, 40–70, 1 year 
Sensitivity analyses: 
1. Adjustment lifetables including different relative risks for different intervals after Hodgkin lymphoma treatment FIT20, 45–70, 2 years Unchanged Unchanged 
2. Adjustment in the lifetables based on other Anderson et al Unchanged FIT47, 45–70, 1 year FIT20, 40–70, 1 year 
3. Higher colorectal cancer risk caused by higher adenoma onset FIT47, 40–70, 1 year FIT47, 45–70, 1 year Unchanged 
4. Different adenoma localization Unchanged Unchanged Unchanged 
5. 1.33-fold higher colorectal cancer relative survival Unchanged Unchanged Unchanged 
6. Systematic FIT negative results Unchanged Unchanged Unchanged 
7. FIT sensitivity for medium adenomas (6–9 mm) as general population Unchanged Unchanged FIT10, 40–70, 2 years 
8. Colorectal cancer risk caused directly by a combination of higher adenoma onset and faster adenoma progression Unchanged Unchanged FIT20, 40–70, 1 year 
9. Lower complete colonoscopy examination rate (92%) Unchanged Unchanged Unchanged 
10. Higher colorectal cancer treatment costs (+50%) FIT10, 45–70, 2 years FIT20, 45–70, 2 years FIT10, 40–70, 2 years 
11. Higher colorectal cancer treatment costs (+100%) FIT10, 45–70, 2 years FIT10, 45–70, 2 years FIT20, 40–70, 1 year 
12. Discounting factor (4% costs, 1.5% benefits) FIT47, 40–70, 1 year FIT47, 45–70, 1 year FIT20, 40–70, 1 year 

*Assuming a willingness to pay threshold of €20,000 per LYG from surveillance.

Recent studies have suggested that Hodgkin lymphoma survivors, who received IRT, procarbazine-containing chemotherapy or both, should undergo colorectal cancer surveillance at an earlier age than recommended in population screening programs due to their increased risk of developing colorectal cancer before age 55 (1–4) and the high prevalence of colonic advanced neoplasia already at young ages (6). Using an established micro-simulation model, we found that FIT is the most cost-effective colorectal cancer surveillance strategy in this population, regardless of the Hodgkin lymphoma treatment associated colorectal cancer risk. Depending on the Hodgkin lymphoma treatment, the optimal age of commencing surveillance ranged from 40 to 45 years, which is earlier than practiced in most colorectal cancer screening programs. We showed that the optimal FIT positivity cutoff was 47 μg Hb/g feces when offering FIT annually to Hodgkin lymphoma survivors in general (the entire cohort) or to those treated with IRT and procarbazine. This FIT positivity cutoff is also used in the Dutch colorectal cancer screening programme for the average-risk population where, however, FIT is offered biennially. For those Hodgkin lymphoma survivors treated with procarbazine without IRT (patients at lower colorectal cancer risk than those with additional IRT) the same program as for the general population would be beneficial, only with a starting age at 45 (biennially and FIT positivity cutoff (47 μg Hb/g feces).

The earlier optimal age of surveillance invitation reflects the higher risk of colorectal cancer among Hodgkin lymphoma survivors at a younger age (1, 2, 6, 31) and is in line with the increased risk after 10 to over 30 years from Hodgkin lymphoma treatment (at a median age of 27 years; ref. 1). Moreover, our model shows that surveillance for Hodgkin lymphoma survivors could stop at age 70 years, 5 years earlier than recommended in most (European) colorectal cancer screening programs (75 years). This can be related to the high all-cause mortality observed among Hodgkin lymphoma survivors (7, 26). In Hodgkin lymphoma survivors, performing surveillance at an older age might directly result in colorectal cancer overdiagnosis and overtreatment with no improvement of life expectancy (no colorectal cancer death averted).

In line with previous studies on the cost-effectiveness of mt-sDNA in the asymptomatic population (32–35), we found that mt-sDNA was not cost-effective compared with other modalities. Although mt-sDNA was estimated to reduce colorectal cancer incidence and mortality, it was an inefficient surveillance option (less effective and higher costs; ref. 32).

This study has several limitations. Firstly, the MISCAN model assumes that all colorectal cancers arise through a traditional adenoma–carcinoma sequence, and the pathway of development of serrated polyps is not (yet) included in MISCAN. To avoid bias towards FIT surveillance, which is less sensitive for serrated lesions than mt-sDNA (9, 17, 36), we have modelled advanced serrated lesions as large adenomas, assuming the same progression rate for both types of lesions. Previously, in the prospective colonoscopy study, we detected significantly more advanced serrated polyps and serrated polyposis syndrome in Hodgkin lymphoma survivors compared with the general population (6), which may have impact on our model adjustments even though the colorectal cancer risk was considered. As we could not disentangle those colonoscopy results to correctly inform our current model structure, we decided to consider the advanced serrated lesions as advanced adenomas. With this assumption, we could compute the adenoma prevalence in Hodgkin lymphoma survivors in way which allow us to validate the model (Supplementary Fig. S2). Furthermore, by applying FIT parameters computed from data which include both adenomas and serrated lesions in Hodgkin lymphoma survivors, our model results were indirectly adjusted to account the potential presence of serrated lesions. A second limitation is that we cannot inform sensitivity of the FIT and mt-sDNA for colorectal cancer based on the prospective data in Hodgkin lymphoma survivors, because no colorectal cancer was detected in this cohort, only precursor lesions (6, 9). Thus, colorectal cancer sensitivity model parameters were based on data from the average-risk population (29). Furthermore, even for precursor lesions, the sample size was small for evaluating the stool test sensitivity (9). Moreover, the exact pathogenesis of colorectal cancer in Hodgkin lymphoma survivors remains unknown. Previous research by our group detected a higher prevalence of microsatellite instability colorectal cancer in Hodgkin lymphoma survivors due to double somatic mutations in mismatch-repair genes (37), suggesting a faster progression from precursor lesions to colorectal cancer (38). Hence, we performed a specific sensitivity analysis assuming faster progression from adenoma to colorectal cancer in Hodgkin lymphoma survivors. We found that the optimal surveillance strategy was not sensitive to this assumption. Furthermore, we assumed full adherence to follow-up and surveillance procedures because this provides unbiased estimates for optimal surveillance strategies. Results should therefore be used to guide policy, but not to take these results as the estimated impact of that policy. In practice, adherence to surveillance is usually lower, resulting in a lower impact of surveillance than suggested. When imperfect adherence is assumed, this would result in strategies with short intervals and larger age target to compensate for the suboptimal surveillance. This would result in Hodgkin lymphoma survivors who adhere to surveillance to be over-screened. If only 41% of the population would participate, the costs, benefits, and harms of the program would decrease proportionally. Considering the low uptake of colonoscopy screening observed among childhood cancer survivors in US (i.e., 11.5%; ref. 39), our findings about the benefits of FIT surveillance in Hodgkin lymphoma may have vital importance as stool tests are generally characterized by higher participation rates compared with colonoscopy (at least in the average-risk population; ref. 40). To assess the robustness of our modelling estimates, a full probabilistic sensitivity analysis was not performed as extremely resource-demanding. Thus, we focused our assessments carrying out several one-way sensitivity analyses on key specific model parameters. Finally, Hodgkin lymphoma treatment regimens have changed over the past decades with a reduction of radiotherapy volumes and doses and changes in chemotherapy regimens, although procarbazine is still used in e.g., the BEACOPP regimen (41). Hence, patients currently diagnosed with Hodgkin lymphoma might have a lower colorectal cancer risk and a less intensive surveillance could be optimal (2).

One of the strengths of this study is that this is the first cost-effectiveness analysis of stool testing performed for Hodgkin lymphoma survivors. Our study suggests that FIT stool tests are cost-effective modalities for colorectal cancer surveillance in this known high-risk group for developing colorectal cancer. FIT is easy to perform and noninvasive. Reducing the use of colonoscopy surveillance will reduce potential harms (i.e., colonoscopy burden and complications) and be beneficial for the national healthcare system limiting the demand of colonoscopy and the workload of gastroenterologists. This will not only impact Hodgkin lymphoma survivors, but also other high-risk groups. Currently, ongoing research is aiming to evaluate whether stool test surveillance might also be beneficial in other high-risk groups (42).

Colorectal cancer surveillance in Hodgkin lymphoma survivors at increased risk for colorectal cancer (treated with IRT and/or procarbazine-containing chemotherapy) is cost-effective and should commence earlier than in the general population. For all examined Hodgkin lymphoma subgroups, FIT surveillance was the most cost-effective strategy. This implies introduction of surveillance with a modality that is currently not used for surveillance in high-risk groups but is extensively used in population-based colorectal cancer screening programs.

M.C.W. Spaander reports other support from Sysmex, Sentinel; and other support from Norgine outside the submitted work. L.M.G. Moons reports Consultant for Boston Scientific. P.J. Lugtenburg reports grants from Takeda, Servier; personal fees from Celgene, Genmab, Roche, AbbVie, Incyte; and personal fees from Regeneron outside the submitted work. G.A. Meijer reports nonfinancial support from Exact Sciences, Sysmex, Sentinel Ch.SpA, Personal Genome Diagnostics (PGDX), Hartwig Medical Foundation; grants from CZ (OWM Centrale Zorgverzekeraars groep Zorgverzekeraar u.a); and nonfinancial support from DELFi outside the submitted work; in addition, G.A. Meijer has a patent for Several pending; and GA Meijer is cofounder and board member (CSO) of colorectal cancer bioscreen BV. P. Snaebjornsson reports other support from MSD, Bayer; and other support from MEDtalks outside the submitted work. B. Carvalho reports a patent for Protein biomarkers for detection of colorectal cancer licensed to NL 2008707;EP13720130.7;14/396,522;NL 2010276;PCT/NL13/50316;15/444,679;EP19201973.5, a patent for Protein biomarkers (II) for detection of colorectal cancer in stool licensed to 17172531.0;2017-009-02;2017-009-03;2017-009-04;2017-009-05;2017-009-06, and a patent for Progression markers for colorectal cancer licensed to EP19187894.1;PCT/NL2020/050482. M.E. van Leerdam reports grants from the Dutch Society of Gastroenterology and Hepatology (Maag Lever Darm Stichting (MLDS) funding project FP14-04 outside the submitted work. No disclosures were reported by the other authors.

B.L.M. Ykema: Conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration. A. Gini: Conceptualization, resources, data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration. L.S. Rigter: Conceptualization, data curation, methodology, writing–review and editing. M.C.W. Spaander: Conceptualization, investigation, writing–review and editing. L.M.G. Moons: Conceptualization, visualization, writing–review and editing. T.M. Bisseling: Conceptualization, investigation, writing–review and editing. J.P. de Boer: Investigation, writing–review and editing. W.H.M. Verbeek: Investigation, writing–review and editing. P.J. Lugtenburg: Investigation, writing–review and editing. C.P.M. Janus: Investigation, writing–review and editing. E.J. Petersen: Investigation, writing–review and editing. J.M. Roesink: Investigation, writing–review and editing. R.W.M. van der Maazen: Investigation, writing–review and editing. B.M.P. Aleman: Conceptualization, investigation, writing–review and editing. G.A. Meijer: Investigation, writing–review and editing. F.E. Van Leeuwen: Conceptualization, methodology, writing–review and editing. P. Snaebjornsson: Conceptualization, methodology, writing–review and editing. B. Carvalho: Conceptualization, methodology, writing–review and editing. M.E. van Leerdam: Conceptualization, resources, data curation, supervision, validation, investigation, methodology, writing–original draft. I. Lansdorp-Vogelaar: Conceptualization, resources, data curation, software, supervision, validation, investigation, methodology, writing–original draft.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Epidemiology, Biomarkers & Prevention Online (http://cebp.aacrjournals.org/).

1.
van Eggermond
AM
,
Schaapveld
M
,
Janus
CP
,
de Boer
JP
,
Krol
AD
,
Zijlstra
JM
, et al
.
Infradiaphragmatic irradiation and high procarbazine doses increase colorectal cancer risk in Hodgkin lymphoma survivors
.
Br J Cancer
2017
;
117
:
306
14
.
2.
Schaapveld
M
,
Aleman
BM
,
van Eggermond
AM
,
Janus
CP
,
Krol
AD
,
van der Maazen
RW
, et al
.
Second cancer risk up to 40 years after treatment for Hodgkin's lymphoma
.
N Engl J Med
2015
;
373
:
2499
511
.
3.
Swerdlow
AJ
,
Higgins
CD
,
Smith
P
,
Cunningham
D
,
Hancock
BW
,
Horwich
A
, et al
.
Second cancer risk after chemotherapy for Hodgkin's lymphoma: a collaborative British cohort study
.
J Clin Oncol
2011
;
29
:
4096
104
.
4.
Hodgson
DC
,
Gilbert
ES
,
Dores
GM
,
Schonfeld
SJ
,
Lynch
CF
,
Storm
H
, et al
.
Long-term solid cancer risk among 5-year survivors of Hodgkin's lymphoma
.
J Clin Oncol
2007
;
25
:
1489
97
.
5.
Stoop
EM
,
de Haan
MC
,
de Wijkerslooth
TR
,
Bossuyt
PM
,
van Ballegooijen
M
,
Nio
CY
, et al
.
Participation and yield of colonoscopy versus non-cathartic CT colonography in population-based screening for colorectal cancer: a randomized controlled trial
.
Lancet Oncol
2012
;
13
:
55
64
.
6.
Rigter
LS
,
Spaander
MCW
,
Aleman
BMP
,
Bisseling
TM
,
Moons
LM
,
Cats
A
, et al
.
High prevalence of advanced colorectal neoplasia and serrated polyposis syndrome in Hodgkin lymphoma survivors
.
Cancer
2019
;
125
:
990
9
.
7.
de Vries
S
,
Schaapveld
M
,
Janus
CPM
,
Daniels
LA
,
Petersen
EJ
,
van der Maazen
RWM
, et al
.
Long-term cause-specific mortality in Hodgkin lymphoma patients
.
J Natl Cancer Inst
2021
;
113
:
760
9
.
8.
Gini
A
,
Meester
RGS
,
Keshavarz
H
,
Oeffinger
KC
,
Ahmed
S
,
Hodgson
DC
, et al
.
Cost-effectiveness of colonoscopy-based colorectal cancer screening in childhood cancer survivors
.
J Natl Cancer Inst
2019
;
111
:
1161
9
.
9.
Ykema
B
,
Rigter
L
,
Spaander
M
,
Moons
L
,
Bisseling
T
,
Aleman
B
, et al
.
Diagnostic accuracy of stool tests for colorectal cancer surveillance in Hodgkin lymphoma survivors
.
J Clin Med
2020
;
9
:
190
.
10.
van Balegooijen
M
,
Boer
R
,
Habbema
JF
,
Loeve
F
,
van Oortmarssen
GJ
,
Vogelaar
I
, et al
.
Model profiler of the MISCAN-Colon miscosimulation model for colorectal cancer. Deparment of Public Health
,
Erasmus MC
2004
;
Available from
: http://cisnet.flexkb.net/mp/pub/cisnet_colorectal_sloankettering_profile.pdf.
11.
Gini
A
,
Zauber
AG
,
Cenin
DR
,
Omidvari
AH
,
Hempstead
SE
,
Fink
AK
, et al
.
Cost-effectiveness of screening individuals with cystic fibrosis for colorectal cancer
.
Gastroenterology
2018
;
154
:
556
67
.
12.
Knudsen
AB
,
Zauber
AG
,
Rutter
CM
,
Naber
SK
,
Doria-Rose
VP
,
Pabiniak
C
, et al
.
Estimation of benefits, burden, and harms of colorectal cancer screening strategies: modeling study for the US Preventive Services Task Force
.
JAMA
2016
;
315
:
2595
609
.
13.
Lansdorp-Vogelaar
I
,
Goede
SL
,
Bosch
LJW
,
Melotte
V
,
Carvalho
B
,
van Engeland
M
, et al
.
Cost-effectiveness of high-performance biomarker tests vs fecal immunochemical test for noninvasive colorectal cancer screening
.
Clin Gastroenterol Hepatol
2018
;
16
:
504
12
.
14.
Morson
B
.
President's address. The polyp-cancer sequence in the large bowel
.
Proc R Soc Med
1974
;
67
:
451
7
.
15.
Vogelstein
B
,
Fearon
ER
,
Hamilton
SR
,
Kern
SE
,
Preisinger
AC
,
Leppert
M
, et al
.
Genetic alterations during colorectal tumor development
.
N Engl J Med
1988
;
319
:
525
32
.
16.
Holmqvist
AS
,
Chen
Y
,
Berano Teh
J
,
Sun
C
,
Birch
JM
,
van den Bos
C
, et al
.
Risk of solid subsequent malignant neoplasms after childhood Hodgkin lymphoma-Identification of high-risk populations to guide surveillance: a report from the Late Effects Study Group
.
Cancer
2019
;
125
:
1373
83
.
17.
Imperiale
TF
,
Ransohoff
DF
,
Itzkowitz
SH
,
Levin
TR
,
Lavin
P
,
Lidgard
GP
, et al
.
Multitarget stool DNA testing for colorectal-cancer screening
.
N Engl J Med
2014
;
370
:
1287
97
.
18.
RIVM
.
National monitoring and evaluation bowel cancer population screening 2014–2017.
2019. Available from: https://www.rivm.nl/documenten/monitoring-evaluatie-bvo-darmkanker-2017.
19.
Cologuard
.
FAQs.
Available from: https://www.cologuardtest.com/faq/cost.
20.
Lew
JB
,
St John
DJB
,
Macrae
FA
,
Emery
JD
,
Ee
HC
,
Jenkins
MA
, et al
.
Evaluation of the benefits, harms and cost-effectiveness of potential alternatives to iFOBT testing for colorectal cancer screening in Australia
.
Int J Cancer
2018
;
143
:
269
82
.
21.
van der Meulen
MP
,
Lansdorp-Vogelaar
I
,
Goede
SL
,
Kuipers
EJ
,
Dekker
E
,
Stoker
J
, et al
.
Colorectal cancer: cost-effectiveness of colonoscopy versus CT colonography screening with participation rates and costs
.
Radiology
2018
;
287
:
901
11
.
23.
Mark
DH
.
Visualizing cost-effectiveness analysis
.
JAMA
2002
;
287
:
2428
9
.
24.
van den Berg
M
,
de Wit
GA
,
Vijgen
SM
,
Busch
MC
,
Schuit
AJ
. [
Cost-effectiveness of prevention: opportunities for public health policy in the Netherlands
].
Ned Tijdschr Geneeskd
2008
;
152
:
1329
34
.
25.
Mertens
AC
,
Liu
Q
,
Neglia
JP
,
Wasilewski
K
,
Leisenring
W
,
Armstrong
GT
, et al
.
Cause-specific late mortality among 5-year survivors of childhood cancer: the Childhood Cancer Survivor Study
.
J Natl Cancer Inst
2008
;
100
:
1368
79
.
26.
Anderson
C
,
Lund
JL
,
Weaver
MA
,
W
WA
,
Olshan
AF
,
Nichols
HB
.
Noncancer mortality among adolescents and young adults with cancer
.
Cancer
2019
;
125
:
2107
14
.
27.
Rigter
LS
,
Schaapveld
M
,
Janus
CPM
,
Krol
ADG
,
van der Maazen
RWM
,
Roesink
J
, et al
.
Overall and disease-specific survival of Hodgkin lymphoma survivors who subsequently developed gastrointestinal cancer
.
Cancer Med
2019
;
8
:
190
9
.
28.
van der Meulen
MP
,
Lansdorp-Vogelaar
I
,
van Heijningen
EM
,
Kuipers
EJ
,
van Ballegooijen
M
.
Nonbleeding adenomas: evidence of systematic false-negative fecal immunochemical test results and their implications for screening effectiveness—a modeling study
.
Cancer
2016
;
122
:
1680
8
.
29.
Goede
SL
,
van Roon
AH
,
Reijerink
JC
,
van Vuuren
AJ
,
Lansdorp-Vogelaar
I
,
Habbema
JD
, et al
.
Cost-effectiveness of one versus two sample fecal immunochemical testing for colorectal cancer screening
.
Gut
2013
;
62
:
727
34
.
30.
National Health Care Institute
.
Guideline for economic evaluations in healthcare
.
Diemen
;
2016
. https://english.zorginstituutnederland.nl/publications/reports/2016/06/16/guideline-for-economic-evaluations-in-healthcare.
31.
Lieberman
DA
,
Rex
DK
,
Winawer
SJ
,
Giardiello
FM
,
Johnson
DA
,
Levin
TR
.
Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer
.
Gastroenterology
2012
;
143
:
844
57
.
32.
Naber
SK
,
Knudsen
AB
,
Zauber
AG
,
Rutter
CM
,
Fischer
SE
,
Pabiniak
CJ
, et al
.
Cost-effectiveness of a multitarget stool DNA test for colorectal cancer screening of Medicare beneficiaries
.
PLoS One
2019
;
14
:
e0220234
.
33.
Ladabaum
U
,
Mannalithara
A
.
Comparative effectiveness and cost-effectiveness of a multitarget stool DNA test to screen for colorectal neoplasia
.
Gastroenterology
2016
;
151
:
427
39
.
34.
Carethers
JM
.
Fecal DNA testing for colorectal cancer screening
.
Annu Rev Med
2020
;
71
:
59
69
.
35.
Peterse
EFP
,
Meester
RGS
,
de Jonge
L
,
Omidvari
AH
,
Alarid-Escudero
F
,
Knudsen
AB
, et al
.
Comparing the cost-effectiveness of innovative colorectal cancer screening tests
.
J Natl Cancer Inst
2021
;
113
:
154
61
.
36.
Heigh
RI
,
Yab
TC
,
Taylor
WR
,
Hussain
FT
,
Smyrk
TC
,
Mahoney
DW
, et al
.
Detection of colorectal serrated polyps by stool DNA testing: comparison with fecal immunochemical testing for occult blood (FIT)
.
PLoS One
2014
;
9
:
e85659
.
37.
Rigter
LS
,
Snaebjornsson
P
,
Rosenberg
EH
,
Atmodimedjo
PN
,
Aleman
BM
,
Ten Hoeve
J
, et al
.
Double somatic mutations in mismatch repair genes are frequent in colorectal cancer after Hodgkin's lymphoma treatment
.
Gut
2018
;
67
:
447
55
.
38.
Kahi
CJ
.
Screening relevance of sessile serrated polyps
.
Clin Endosc
2019
;
52
:
235
8
.
39.
Nathan
PC
,
Ness
KK
,
Mahoney
MC
,
Li
Z
,
Hudson
MM
,
Ford
JS
, et al
.
Screening and surveillance for second malignant neoplasms in adult survivors of childhood cancer: a report from the childhood cancer survivor study
.
Ann Intern Med
2010
;
153
:
442
51
.
40.
Hol
L
,
van Leerdam
ME
,
van Ballegooijen
M
,
van Vuuren
AJ
,
van Dekken
H
,
Reijerink
JC
, et al
.
Screening for colorectal cancer: randomized trial comparing guaiac-based and immunochemical fecal occult blood testing and flexible sigmoidoscopy
.
Gut
2010
;
59
:
62
8
.
41.
Engert
A
,
Goergen
H
,
Markova
J
,
Pabst
T
,
Meissner
J
,
Zijlstra
JM
, et al
.
Reduced-intensity chemotherapy in patients with advanced-stage Hodgkin lymphoma: updated results of the open-label, international, randomized phase III HD15 Trial by the German Hodgkin Study Group
.
Hemasphere
2017
;
1
:
e5
.
42.
van Lanschot
MC
,
Carvalho
B
,
Coupe
VM
,
van Engeland
M
,
Dekker
E
,
Meijer
GA
.
Molecular stool testing as an alternative for surveillance colonoscopy: a cross-sectional cohort study
.
BMC Cancer
2017
;
17
:
116
.
43.
Human Mortality Database
.
University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available from:
www.mortality.org or www.humanmortality.de.
44.
van Rijn
JC
,
Reitsma
JB
,
Stoker
J
,
Bossuyt
PM
,
van Deventer
SJ
,
Dekker
E
.
Polyp miss rate determined by tandem colonoscopy: a systematic review
.
Am J Gastroenterol
2006
;
101
:
343
50
.
45.
Schroy
PC
3rd
,
Coe
A
,
Chen
CA
,
O'Brien
MJ
,
Heeren
TC
.
Prevalence of advanced colorectal neoplasia in white and black patients undergoing screening colonoscopy in a safety-net hospital
.
Ann Intern Med
2013
;
159
:
13
20
.
46.
Warren
JL
,
Klabunde
CN
,
Mariotto
AB
,
Meekins
A
,
Topor
M
,
Brown
ML
, et al
.
Adverse events after outpatient colonoscopy in the Medicare population
.
Ann Intern Med
2009
;
150
:
849
57
.
47.
Gatto
NM
,
Frucht
H
,
Sundararajan
V
,
Jacobson
JS
,
Grann
VR
,
Neugut
AI
.
Risk of perforation after colonoscopy and sigmoidoscopy: a population-based study
.
J Natl Cancer Inst
2003
;
95
:
230
6
.

Supplementary data