Treatment Beyond Progression After Anti-PD-1 Blockade in Hepatocellular Carcinoma

Immune checkpoint inhibitors (ICI) can induce atypical tumor responses including pseudoprogression in a subset of patients who may benefit from treatment beyond progression. While ICIs have emerged as frontline treatments for hepatocellular carcinoma (HCC) and are associated with clinical benefit in a minority of patients, it is unclear whether treatment beyond progression has utility in this disease type. In a multicenter cohort analysis, treatment beyond progression was associated with no new safety signals, objective responses in 5.8% of patients, and disease control in 44% of patients. Progression-free survival and overall survival were comparable between patients treated beyond progression and patients treated with subsequent therapies, demonstrating that treatment beyond progression was not detrimental to survival outcomes. Rather, treatment beyond progression may benefit select patients with HCC and could represent a viable strategy for maximizing treatment benefit in these patients. Significance: Treatment beyond progression with ICIs in patients with HCC is safe and may benefit a subset of patients due to later-onset tumor responses or disease stability. These findings may guide the design of trials testing ICIs in HCC and the use of treatment beyond progression in routine practice.


Introduction
Hepatocellular carcinoma (HCC) is the most prevalent primary liver cancer with a rising incidence in many countries and is a leading cause of cancer deaths worldwide (1).Immune checkpoint inhibitors (ICI) are currently the preferred frontline agents for the treatment of advanced HCC and are associated with prolonged overall survival (OS2) benefit and tumor responses compared with tyrosine kinase inhibitors (2).However, determining the efficacy of ICIs in clinical practice is hampered by evidence that ICIs can induce immune-related responses in tumor lesions that may not be accurately classified using standard criteria including RECIST 1.1 (3).RECIST 1.1 defines unequivocal disease progression as an increase in the dimensions of tumor lesions by at least 20% or the observation of new tumor lesions, and this is conventionally applied in clinical trial endpoints and guides treatment continuation in routine practice (3).
Nonetheless, immune-related treatment responses include delayed antitumor

Data Collection
Patient characteristics, treatment history, and toxicities were manually abstracted from medical records including clinic notes, telephone encounters, radiology images and reports, laboratory results, medication lists, and hospitalization records.Radiographic responses were based on RECIST 1.1.Progression-free survival (PFS2) after initial progression to ICIs was calculated from the time of first progression to ICIs to subsequent progression of disease or death.OS2 was calculated from the time of first progression to ICIs to death.

Statistical Analysis
The χ 2 or two-sided Mann-Whitney U test was used to assess differences in the frequency of categorical characteristics between patient cohorts.Median survival times were calculated using the Kaplan-Meier method and the logrank test was used to assess differences in survival.Data analysis was performed using SPSS 24 (IBM).

Results
Among 144 patients with advanced HCC treated with anti-PD-1/L1 therapies, 102 patients had progressive disease at the data cut-off date according to RECIST 1.1, and 68 patients received further treatment (34 treated beyond progression and 34 treated with other subsequent therapies).Clinical characteristics were not significantly different between patients who received treatment beyond progression and patients who received subsequent line therapies (Table 1).In contrast, patients who received no other subsequent line therapy including treatment beyond progression were more frequently associated with a poor baseline Eastern Cooperative Oncology Group (ECOG) status (Table 1).Nearly half of the patients treated beyond progression were initially treated with nivolumab (Table 2).PFS to the first occurrence of progression after initial ICI treatment was shorter among patients treated beyond progression (2.7 months) and patients who did not receive subsequent treatment (2.3 months) compared with patients who received subsequent line therapies (4.5 months; log-rank test,  and 11 patients treated with subsequent line therapy (9 of which increased from 0 to 1, and 2 of which increased from 0 or 1 to 2).
Among patients treated beyond progression, objective tumor responses according to RECIST 1.1 including partial and complete responses were observed in 5.8% (1 partial response, 1 complete response), while stable disease was observed in 38% (Fig. 1).A total of 11.8% of patients demonstrated evidence of tumor shrinkage after initial progression of disease (Fig. 1).Among the 2 patients with objective responses after treatment beyond progression, 1 patient developed new hepatic lesions 10 months after initiation of nivolumab, which subsequently regressed (Fig. 2).The other patient had a 25% growth of target lesions after starting atezolizumab plus bevacizumab, which diminished at the next restaging scan with further treatment of the same regimen (Fig. 2).
To assess survival outcomes after initial progression, we calculated PFS2 and OS2 across patient groups where time to events was calculated from the time of initial progression.The median PFS2 of patients treated beyond progression and treated with subsequent line therapy was 3.7 and 3.2 months, respectively (log-rank test, P = 0.81; Fig. 3A).The median OS2 of patients treated beyond progression and treated with non-ICI subsequent line therapy was 17 and 13 months, respectively (log-rank test, P = 0.11; Fig. 3B).Two-month land-mark analyses showed that disease control defined as complete or partial responses and stable disease after treatment beyond progression was associated with prolonged PFS2 (7.9 vs. 3 months; log-rank test, P < 0.001) and OS2 (17 vs. 14.8 months; log-rank test, P = 0.45; Fig. 3C and D).This suggests that subsequent tumor responses in patients treated beyond progression may be predictive of clinical benefit in select patients.

Subset analyses of patients with viral or non-viral etiologies of cirrhosis and the
presence and absence of extrahepatic disease were performed to assess whether clinical factors may be associated with benefit after treatment beyond progression.Given low objective response rates in patients treated beyond progression, we compared disease control rates (complete response/partial response/stable disease) between groups as well as median PFS2.There was no statistical difference in disease control rates between viral versus non-viral etiologies (36% vs. Treatment adverse events occurred in 14.7% of patients during treatment beyond progression, all of which were grade 1 or 2. No patient permanently discontinued ICIs during treatment beyond progression due to toxicities.

Discussion
These results indicate that treatment beyond progression with ICIs in HCC is safe and may be a therapeutic strategy in select individuals.Disease control in a substantial proportion of patients as well as objective tumor responses in a subset of patients confirms that delayed antitumor effects in HCC may be observed with ICIs.In addition, the lack of a difference in PFS2 and OS2 between patients treated beyond progression and patients treated with subsequent therapies suggests that continuation of ICIs after initial progression is not detrimental to survival.Most demographic and clinical characteristics were not significantly different between patients who were treated beyond progression and those who received subsequent line therapy in this study, but patients with a shorter time to first ICI progression and without deterioration of ECOG status were more likely to be treated beyond progression.Understanding provider and patient preferences that may influence the use of treatment beyond progression will need to be delineated to implement such treatment strategies.
Further research is necessary to explore the generalizability of these results because the utilization of treatment beyond progression may vary across institutions.Review of imaging was not blinded to clinical treatment decisions, which was a limitation of this study.Because of the retrospective nature of this study, it was difficult to capture specific reasons why treatment was continued beyond progression.Although the relatively small size of our cohort limits the generalizability of our findings and the ability to perform a matched cohort analysis, this study is among the first to evaluate outcomes of treatment beyond progression to anti-PD-(L)1 therapy in HCC.Larger-scale studies are warranted to provide insight into which subset of patients with HCC would benefit from treatment beyond progression and the appropriate duration of such treatment.
Nonetheless, our results are aligned with findings from several studies evaluating the continuation of ICIs beyond progression in non-small cell lung cancer (NSCLC), renal cell carcinoma, and melanoma, where a majority of patients demonstrated stable disease without new safety signals (5)(6)(7).Meta-analyses in melanoma and NSCLC also indicate that treatment beyond progression may be associated with improved survival outcomes, but prospective evidence is lacking (5, 7).However, even in these other cancer types, it remains unclear whether treatment beyond progression is beneficial for unselected patients, and if not, whether clinical or biological factors may guide treatment selection.
Further research identifying biomarkers that predict outcomes after treatment beyond progression in HCC will be needed to guide patient selection for the continuation of ICIs.In addition, whether modified RECIST proposed for other malignancies may lead to more accurate assessments of ICI benefit in HCC Lim et al.
remains to be clarified because our results suggest that pseudoprogression with ICIs in HCC is rare (8)(9)(10).
In summary, treatment beyond progression with ICIs may benefit a subset of patients with HCC due to the potential for later-onset tumor responses or disease stability.Given that treatment beyond progression may delay the use of subsequent line therapies, including oral kinase inhibitors which are associated with a less tolerable toxicity profile, treatment beyond progression with ICIs represents a promising avenue for maximizing treatment benefit in patients with HCC (11).

FIGURE 1
FIGURE 1 Tumor response after treatment beyond progression in HCC.Waterfall plot of the maximal tumor reduction after treatment beyond progression with ICIs.Dashed columns indicate patients with progressive disease due to the appearance of new lesions.

43%,
Fisher exact test P value = 0.53) and the presence versus absence of extrahepatic disease (47% vs. 33%, Fisher exact test P value = 0.52).Similarly, there was no statistical difference in the median PFS2 between viral versus non-viral etiologies (3.7 vs. 3 months, log-rank P value = 0.74) and the presence versus absence of extrahepatic disease (3.7 vs. 3.5 months, log-rank P value = 0.8).

FIGURE 2
FIGURE 2 Changes in tumor measurements among patients with objective responses after treatment beyond progression.Spider plots indicate changes in tumor dimensions in 2 patients with progressive disease according to RECIST 1.1 followed by response after continuing ICI.

FIGURE 3
FIGURE 3 PFS2 and OS2 after treatment beyond progression in HCC.A, Kaplan-Meier curves depict PFS2 among patients treated beyond progression and treated with subsequent therapies.B, Kaplan-Meier curves depict OS2 among patients treated beyond progression and treated with subsequent therapies.C, Two-month landmark analysis of PFS2 among patients treated beyond progression who had evidence of disease control and patients with progressive disease.D, Two-month landmark analysis of OS2 among patients treated beyond progression who had evidence of disease control and patients with progressive disease.

TABLE 1
Characteristics of patients with HCC treated beyond progression, treated with subsequent therapies, and received no subsequent therapies

Treatment beyond progression Treatment with subsequent therapy No subsequent therapy P P
a Statistical significance tested using two-sided Mann-Whitney U test.b Statistical significance tested using the χ 2 test.

TABLE 2
First-line regimens used for patients with HCC treated beyond