Long-term Consequences of Pelvic Irradiation: Toxicities, Challenges, and Therapeutic Opportunities with Pharmacologic Mitigators

Radiotherapy is frequently used in the treatment of pelvic cancers, including those of genitourinary, gynecologic, and anorectal origin. In locally advanced rectal cancer, preoperative chemo-radiotherapy (CRT) followed by radical surgery, is a standard-of-care (1). Pelvic CRT is frequently used in the treatment of cervical cancer (2). For localized prostate cancer, pelvic irradiation is often employed as a definitive or adjuvant treatment (3).

Unfortunately, irradiation of normal tissue within the radiation field is unavoidable. The use of three-dimensional conformal or intensity-modulated radiotherapy (IMRT), along with improvements in image guidance and patient immobilization, have reduced, but not eliminated, the incidence and severity of toxicities. In prostate cancer, modern radiotherapy methods have enabled dose escalation, leading to 95% local tumor control with <3% risk of chronic normal tissue complications (4). Utilization of extended field IMRT for cervix cancer, which included involved pelvic and para-aortic lymph nodes, led to a local control rate of 80%, with a 10% incidence of late grade 3 toxicity (5). Use of a simultaneous integrated boost, using IMRT, in 29 patients with gynecologic tumors led to a local control rate of 70%, with a 20%–30% incidence of grade 1–2 late toxicities, but no reported grade 3/4 toxicities (6). Further dose escalation in both scenarios is possible, but the risk of more severe late complications will also increase. Local control rates for rectal cancer are in the range of 90%–95% for modern conformal radiotherapy with surgery. The incidence of severe late complications is estimated to be between 5% and 10% (7, 8). In summary, dose escalation with conformal methods has improved local tumor control for prostate cancer, but did not increase normal tissue complication rates. In contrast, modern radiotherapy methods have improved local control for rectal and cervix cancer, but they have not lowered the complication rate to the level seen with prostate.

While the relative incidence of pelvic radiation disease is small, one has to consider the total number of patients affected. It is estimated that by 2030, the total number of patient survivors in the United States with irradiated pelvic tumors will be close to 1 million (9). This includes patients with cancers of the prostate (627,000), uterus (164,000), rectum (160,000), anus (49,000), and cervix (71,000). The total number of patients who could develop pelvic radiation disease is notable because of the number of projected survivors. For example, the incidence of chronic pelvic radiation injury for prostate cancer, assuming 627,000 survivors, is estimated to be between 6,300 and 19,000, assuming incidence rates between 1% and 3%. For the remaining cancers discussed above, the incidence of chronic pelvic radiation injury is projected to be 23,000–47,000, assuming a 5%–10% incidence rate. Furthermore, the incidence of these injuries will remain higher in resource limited countries where modern radiotherapy methods are not possible.

Organs-at-risk for chronic pelvic damage include the anus, rectum, prostate, gynecologic organs, bladder, pelvic bones, and small and large bowel. Damage to the small bowel, termed radiation enteropathy, or -enteritis, is the best-studied late effect. Concerns about small bowel toxicity commonly constrain the prescribed dose. Although there have been a number of positive preclinical and clinical studies addressing treatment and prevention of chronic radiation injury, none have been translated to standard of care (10).

This review highlights drugs that could target key cell signaling pathways involved in both acute and chronic radiation injury. We examine both preclinical and clinical trials with relevant drugs and conclude with a discussion of current and potential trials that may reduce incidence or severity of pelvic radiation disease, while not protecting tumor. We focus on the subject of pelvic radiation disease, but the principles apply to chronic toxicities that occur in other irradiated tumor sites, such as lung, head, and neck and brain. The total number of 5 year cancer survivors treated with radiotherapy is estimated to be over 4 million by the year 2030 (9). Many of these survivors of diseases outside the pelvis are also at risk for normal tissue injury that will reduce quality of life.

Development of injury to tissues after radiotherapy follows sequential effects that begin with damage manifestation in rapidly proliferating cell populations, such as mucosa. Mucosal changes after radiotherapy can be seen within hours. In the small bowel, death of intestinal stem cells results in insufficient replacement of villus epithelium, breakdown of the mucosal barrier, and mucosal inflammation (10). Fluid and nutrient absorption is impaired and mucosal ulcerations develop, leading to diarrhea, nausea, abdominal pain, mucous discharge, and rectal bleeding. If ulcerations persist, bacterial translocation can occur and may lead to systemic inflammation. Acute symptoms develop in 60%–80% of patients during pelvic radiotherapy. These symptoms usually respond to medical treatment and resolve within 1–3 months after radiotherapy (10, 11).

For the majority of patients, the acute effects are self-limiting and the patients will not suffer significant long-term consequences.

For the subset of patients who develop chronic pelvic radiation disease, a persistent unchecked inflammatory response appears to be causative. A discussion of this pathophysiology follows, with emphasis on potential therapeutic targets. The factors that contribute to long-term tissue damage start with a “danger response” shortly after being irradiated; the tissue recognizes that it has been injured and initiates a non-self-limiting wound-healing response (10, 12). The following discussion follows the chronologic steps leading to chronic tissue injury post radiotherapy.

In normal wound healing, vessels upregulate adhesion molecules to facilitate extravasation of immune cells, such as neutrophils and macrophages, into the extravascular space (13). These same effects occur very soon after radiotherapy. Leukocyte–endothelial cell interactions are increased in subcutaneous microvasculature within minutes to hours after exposure to 6 Gy (14). These very early effects were blocked by an inhibitor of platelet activating factor (PAF). PAF upregulates selectins, which are responsible for leukocyte rolling on endothelium. Leukocyte–endothelial cell interactions are also upregulated within 2 hours of exposure to 20 Gy in the mesenteric vasculature (15). Hyperpermeability to albumin was also observed in the mesentery, up to 6 hours after radiation exposure. Inhibition of intracellular adhesion molecule (ICAM)-1, the adhesion molecule responsible for leukocyte adhesion to endothelial cells, blocked the hyperpermeability response to radiotherapy. However, increases in leukocyte–endothelial cell response and hyperpermeability can persist for days to weeks after radiotherapy (16). In the brain, leukocyte–endothelial cell interactions were upregulated up to 3 weeks after single doses of 15, 22.5, and 30 Gy (17). At 3 weeks post radiotherapy, microthrombi and aneurysms were observed, which indicated that microvascular damage could occur in the brain as early as 3 weeks after radiotherapy. These results strongly suggest that leukocyte–endothelial cell interactions participate in the initial wound-healing response post radiotherapy, but may also contribute to the prolongation of the inflammatory and tissue fibrotic response. There have not been any publications to indicate whether amelioration of acute effects with PAF inhibitors or antibodies to integrins have any effects on chronic radiotherapy injury.

The hyperpermeability of injured vessels facilitates extravasation of fibrinogen into the extravascular space (18). Fibrinogen undergoes fibrinopeptide cleavage by thrombin, which leads to subsequent crosslinking of fibrin by enzymes, such as tissue transglutaminase (19). Crosslinking of tissue transglutaminase activates latent TGFβ, which furthers tissue fibrosis by promoting deposition of collagen from tissue fibroblasts (20). Tissue hypoxia and/or increased oxidative stress after radiotherapy may also contribute to the promulgation of unregulated wound healing by upregulation of transcription factors, hypoxia-inducible factor-1 (HIF-1; ref. 21), and NFκB (22–24). These transcription factors are potentially targetable.

Upregulation of HIF-1 transcriptional activity is driven by an increase in the HIF-1α subunit of HIF-1, which is regulated posttranscriptionally by hypoxia and oxidative stress (25). HIF-1 regulates hundreds of genes, many of which could contribute to the wound-healing response (21). For example, HIF-1 upregulates VEGF expression; VEGF potently stimulates hyperpermeability and angiogenesis; both are required for full initiation of the normal wound-healing process (20). Furthermore, virtually all of the enzymes involved in cross-linking of collagen, a necessary step in the development of fibrosis, are HIF-1 target genes (26). Thus, upregulation of HIF-1 post radiotherapy can contribute to a chronic wound-healing response contributing to fibrosis.

NFκB is a rapid acting prosurvival (27), proinflammatory (28), and proimmune (29) transcription factor that controls expression of many proinflammatory cytokines. It is rapidly upregulated by bacterial products, radiation, and oxidative stress, among other factors. Although HIF-1α regulation is primarily posttranscriptional, NFκB controls its transcription rate (30).

Damage to nerves after radiotherapy could also contribute to late tissue damage. Chronic effects of radiotherapy on pelvic nerve function have been studied in a canine model of stereotactic radiotherapy for prostate cancer (31). Reduction in neurofilament content of penile nerves was observed along with axonal degeneration and reduced nerve conduction velocity. Such effects could contribute to erectile dysfunction after radiotherapy. It is also likely that nerve injury is at least partially the product of microvascular damage to vessels that supply these larger nerves (32). Oxidative stress plays a major role in this pathophysiology (33).

Consistent with the chronic wound-healing response discussed above, there is increased fibrosis of the bowel after radiotherapy (34). Fibrosis contributes to: (i) altered intestinal transit, (ii) nutrient malabsorption, (iii) impaired gut motility, (iv) fistulae formation, (v) rectal bleeding, (vi) partial or complete bowel obstruction due to strictures, and (vii) perforation with sepsis (35). These chronic toxicities are not currently preventable or reversible in the clinic. As will be discussed below, mitigators in development could be effective in prevention or reduction in severity of pelvic radiation disease.

Figure 1 summarizes the primary factors that contribute to pelvic radiation disease and the functional consequences of this chronic damage.

Comorbidities may contribute to higher incidence of chronic normal tissue injury post radiotherapy (36–38). A few examples are discussed below. These and others are summarized in Supplementary Table S1.

Eifel and colleagues examined records of nearly 3,500 patients with cervix cancer, treated with definitive radiotherapy between 1960 and 1994 (36). Although these patients were not treated with modern conformal methods, the results revealed important information about patient-related factors that can contribute to chronic injury. Factors that were associated with overall risk included race, body mass index (BMI), smoking history, and prior histories of pelvic inflammatory disease or venereal disease. Patients who smoked > 1 pack per day exhibited an overall 2.3-fold increase in risk of chronic injury in multivariate analysis. The associated incidence of small bowel complications increased from 2% for nonsmokers to 11.1% for patients who smoked more than 1 pack per day. Hispanic race was associated with decreased overall risk, whereas history of pelvic infection increased overall risk of injury by 1.7 fold. Patients with BMI < 22 were at 1.29-fold more likely to experience chronic injury.

Willett and colleagues examined the incidence of acute and chronic bowel injury in patients with rectal cancer with a history of inflammatory bowel disease (37). The 5-year actuarial rate of late toxicity was 73% for patients treated with conventional therapy, but those patients who were treated with approaches to focus more exactly on the tumor volume experienced an actuarial complication rate of 25%. Other patient-related factors were not reported.

Wedlake and colleagues enrolled 308 patients into a study to examine the incidence of pelvic radiation disease at one year post radiotherapy for pelvic malignancies (39). Patient-related factors considered in this analysis included age, BMI, history of previous surgery, smoking history, diabetes, hypertension, history of inflammatory bowel disease, connective tissue disease, regular NSAID use, and whether they also received chemotherapy. 193 patients were eligible for analysis at one year. There was weak association between acute symptom scores and incidence of pelvic radiation disease, but no statistically significant association between any of the patient-related factors and toxicity was observed. The study included patients with any pelvic malignancy. The heterogeneity of treatment received and site of radiotherapy may have contributed to the inability to discern importance of patient-related risk factors.

Huscher and colleagues examined the incidence of chronic small bowel toxicity among 806 patients treated postoperatively for endometrial or cervix cancer (40). Factors associated with chronic injury in univariate analysis included radiotherapy dose per fraction (> vs. <1.8 Gy), patient age (< vs. >60 years) and whether or not the patients developed grade 3–4 acute toxicity. The RRs for these three factors in multivariate analysis were 2.81, 1.02 (continuous variable, per year), and 3.01, respectively.

In summary, although there is heterogeneity in reporting of patient-related factors and probability for pelvic radiation disease, smoking history (36, 39), BMI (36, 39), history of prior inflammatory disease in the irradiated site (36, 37), hypertension (41), and age (40) were linked to risk of pelvic radiation disease. Diabetes (38) and collagen vascular disease (42) have been associated with increased risk for late radiotherapy damage in some reports. In one report, hypertension was associated with lower risk for pelvic radiation disease after prostate radiotherapy (41). Severity of acute symptoms has also been reported to be associated with greater risk for chronic damage (39, 40). However, acute toxicity data could not be used for patient selection a priori. Treatment methods, such as radiotherapy dose/fraction, total radiotherapy dose, addition of chemotherapy, and efforts to reduce radiotherapy dose to key normal tissues are linked to risks for pelvic radiation disease. This article focuses on patient-related factors and therefore, these treatment-related factors will not be discussed further.

Polymorphisms may also contribute to elevated risk for normal tissue injury.

Recently, the C-509T polymorphism (22%–55% incidence, depending on race) of TGFβ was studied prospectively as a secondary endpoint in a series of 184 patients with breast cancer who had been followed for a minimum of 3 years post radiotherapy. In multivariate analyses, the C-509T polymorphism was associated with a 4.47-fold higher risk of developing grade 2 or greater breast fibrosis, compared with the wild-type allele (Table 1; ref. 43). The role that this polymorphism might play in pelvic radiation disease has not been reported.

It has been reported that endothelial plasminogen activator inhibitor-1 (PAI-1) contributes to acute and chronic gastrointestinal injury after radiotherapy to intestine (44). Clinically, genetic polymorphisms of PAI-1 (Table 1) affect the incidence of acute injury after radiation to patients with rectal cancer (45). The rs1050955 variants were associated with lower risk of diarrhea, with ORs of 0.395 for GG versus AA. The rs2227631 variants of PAI-1 were associated with higher risk for anal incontinence with ORs of 2.079 and 3.064 for the AG and GG SNPs, respectively, compared with the AA wild-type. These results strongly support the evaluation of potential mitigators of PAI-1 as a means to protect against acute and chronic intestinal injury after radiation, but caution should be used for patients with rs1050955 variants because they appear to have lower risk for acute effects of radiotherapy.

Oh and colleagues used genome-wide association methods of SNPs to predict for incidence of rectal bleeding and impotence, both of which can occur after prostate radiotherapy (46). They argued that evaluation of individual SNPs may lead to false positive or negative results, because single SNPs may have only a small influence by themselves. The investigators deleted rare SNPs from this analysis, keeping those that have a reasonable prevalence in the population. The analysis was based on a cohort of 368 patients with irradiated prostate cancer who had been enrolled previously into a GWAS analysis protocol. Clinical risk factors were included in these models. Interestingly, there was no overlap in the final identification of 198 and 90 SNPs associated with the development of rectal bleeding or impotence, respectively. Prominent processes associated with rectal bleeding included genes involved in potassium, metal, and calcium ion transport and transmembrane transport. Prominent processes associated with impotence included leukocyte chemotaxis and migration. In a separate report, this same group performed a similar analysis for urinary symptoms after radiotherapy for prostate cancer (47). Kerns and colleagues performed a meta-analysis of genome wide association in 1,576 patients with prostate cancer treated with radiotherapy enrolled into four trials. Prospectively collected symptoms included urinary frequency, decreased urine stream, rectal bleeding, and overall toxicity. These toxicities were examined for SNP associations using multivariate analysis. Two SNPs in noncoding regions were associated with either decreased urine stream or urinary frequency, rs7720298 and rs17599026, respectively. It is challenging to develop therapeutic targets from SNP analyses, but these analyses could assist in identifying patients at highest risk for development of pelvic radiation disease, at least for prostate cancer, where the SNP studies have been focused to date.

Despite the fact that patient comorbidities and SNPs influence the incidence and severity of normal tissue damage, few pre-clinical studies to evaluate impact of these comorbidities on normal tissue damage after therapeutic radiation have been published. Ettarh and colleagues examined acute effects of radiotherapy on small bowel of streptozocin-induced diabetic mice, but found no evidence for differences in damage at this early time point (48). We were unable to find any murine studies where pelvic radiation disease was studied in the context of comorbidities. Murine models of autoimmune disease (49, 50), type I (51), and type II diabetes (52, 53) and inflammatory bowel disease (53, 54) could prove useful for preclinical evaluation of mitigators. Furthermore, genetic engineering could be used to introduce limited numbers of SNPs into mice. Such mice could provide some validation of whether SNPs associated with pelvic radiation disease actually directly contribute to the pathology of pelvic radiation disease.

There are important challenges to translating murine study results to the clinic. First, studies in mice are typically done with young, healthy mice, whereas patients who develop pelvic radiation disease are older. Second, chronic normal tissue injury in mice is not studied in tumor-bearing individuals. The extent to which damage to normal tissue is influenced by the presence of tumor cannot be studied in mice. Third, the symptoms of pelvic radiation disease are difficult to quantify in mice (e.g., decreased urine stream, polyuria, pelvic pain).

A prior challenge for murine studies has been that irradiation of the pelvis could not be done in a manner similar to that used in humans. Large volumes of pelvic tissue had to be irradiated and mortality was traditionally quite high. The availability of high precision micro-CT/irradiators has now made it possible to simulate target localization and precise radiotherapy delivery to target tissues in the pelvis (54).

Tumor bearing companion dogs represent a viable intermediate model between the mouse and humans. Tumor-bearing companion dogs tend to be older and often have concurrent diseases that might contribute to overall pathology (55). Dogs develop a number of pelvic tumors, including prostate cancer (56), bladder cancer (57) and perianal gland tumors (58). They develop late effects that are very similar in scope and severity to what is seen in human patients. In a retrospective study, of dogs treated with radiotherapy to tumors located in the pelvis, incidence of late pelvic damage was reported to be to be nearly 50% (59). Examples of late damage (defined as requiring medical or surgical intervention for symptoms arising >6 months post radiotherapy) included tissue necrosis, chronic colitis, strictures, rectal perforation, bladder wall thickening, pelvic limb edema, and pelvic pain. Canine trials have been conducted previously, to examine pharmacologic mitigators in combination with radiotherapy (60, 61). Most veterinary radiotherapy clinics now have linear accelerators and use modern radiotherapy methods, including IMRT, FLASH, and stereotactic radiotherapy (60). Including tumor-bearing companion canine patients in studies of mitigators may prove useful in a variety of ways, prior to introduction into human trials.

Strategies or interventions for protecting normal tissues from radiation injury may be divided into two conceptual categories (10). The first approach is to directly protect against or mitigate the oxygen free radical injury that occurs during and after radiotherapy. Examples of mitigators of this type include antioxidants, free radical scavengers, and various cytoprotectors (62). The second approach is to target downstream tissue responses from radiation exposure. Examples include immune-modulators (63), enterotrophic drugs (64), compounds that regulate intraluminal contents (65), drugs that inhibit key signal transduction pathways (66), and alteration of fibroblast metabolic radiotherapy response (67). It is not possible to discuss all agents under development in this review. We summarize studies that provide quantitative data on extent of protection for preclinical (Table 2) and clinical studies (Table 3).

Two important issues should be addressed when considering potential mitigators of radiotherapy damage. These two issues are relevant to all drugs: (i) scheduling of mitigators with radiotherapy and (ii) verifying that mitigators do not protect tumor against radiotherapy damage. Examples of these considerations are shown below.

Scheduling of mitigators with radiotherapy depends upon the pharmacologic target and the pharmacokinetics/pharmacodynamics of the mitigator. The classical free radical scavenger, WR2721, had to be administered immediately before radiotherapy, because the targets were the free radicals generated by radiotherapy itself (61). Other agents target downstream tissue reactions to radiotherapy, such as chronic oxidative stress and activation of NFκB. Examples include an isoflavone, genestein, a salen-manganese (Mn) superoxide dismutase-catalase mimetic, and a redox-active Mn-porphyrin. Significant lung radioprotection was observed when genestein or the salen-Mn mimetic was administered concurrently or two weeks after whole thorax irradiation (68). Administration of the redox-active Mn-porphyrin for several days before radiotherapy was shown to achieve therapeutically active concentrations in tissue (69). This schedule of drug administration reduced neurocognitive loss (69, 70), mucositis, and xerostomia after radiotherapy to brain or head and neck (71), respectively. A similar compound partially ameliorated chronic radiation–induced lung damage, even when administered 8 weeks after radiation exposure (72).

Whenever radioprotective agents are used, one has to also determine whether the agents protect tumor. The redox-active Mn-porphyrin has been shown to protect normal tissue, but sensitize tumor to radiotherapy (70, 71). It has been hypothesized that the differences in tumor versus normal tissue response to this agent are related to differences in drug accumulation (much higher in tumors) and the effects of Mn-porphyrin on protein modifications that affect signal transduction downstream. Examples include NFκB, nuclear factor-like 2 (Nrf2), MAPK, and phosphatases (73).

Oxidative stress can promote the wound healing response by increasing HIF-1 transcriptional activity, as well as by directly contributing to tissue damage by oxidation. One of the major contributors to oxidative stress is the macrophage. The M1 phenotype contains activated nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, an enzyme that produces very high fluxes of superoxide (74). The M1 macrophage contributes to inflammation and tissue damage. The M2 macrophage, in contrast, is anti-inflammatory (75). RT increases infiltration of M1 macrophages into irradiated normal tissue in mice, contributing to radiation pathology (44). Isoflavones (75) and the redox-active Mn-porphyrin downregulate M1 infiltration into tissue (70).

Mice with PAI-1 knockout of endothelial cells are less sensitive to RT, and experience reduced severity of late fibrosis (44). The protective effect of PAI-1 knockout is associated with a reduction in M1 macrophages and an increase in M2 content in irradiated bowel. These pre-clinical and clinical results strongly support the evaluation of potential mitigators of the PAI-1 response as a means to protect against acute and chronic intestinal injury after radiation.

However, one needs to insert a cautionary note. Infiltration of M1 macrophages into tumors is a desired outcome after radiation because M1 macrophages contribute to the innate immune response (74). In contrast, M2 macrophages contribute to immune tolerance and promote tumor angiogenesis and metastasis. Thus, the desired balance of M1/M2 after RT is opposite for tumor vs. normal tissue. Future studies should evaluate both M1/M2 populations in tumor and normal tissue RT responses when combined with drugs to ameliorate the PAI-1 response.

HIF-1 levels increase in tumors after radiation and this increase protects tumors from radiation damage (76). The story with irradiated normal tissue may be quite different from tumor. Taniguchi and colleagues reported that inhibition of the prolyl hydroxylases that are responsible for priming HIF-1 and HIF-2 for proteasomal degradation, protect bowel against the acute effects of radiation (77). HIF-2 was found to be the primary transcription factor responsible for gastrointestinal mortality. Dimethyloxalylglycine (DMOG), an oxoglutarate analogue, inhibits the prolyl hydroxylases, stabilizing both HIF-1α and HIF-2α. DMOG exhibited profound radioprotection of acute injury to bowel after radiation exposure via a HIF-2 mediated mechanism. In contrast, DMOG did not protect tumors from RT.

In Table 2, we indicate which agents have been studied for radioprotective effects preclinically in normal tissues and which have also been evaluated for tumor radioprotection. In Table 3, we used clinicaltrials.gov to identify: 1) clinical trials that have tested potential mitigators in combination with RT. 2) other trials where relevant drugs were used, but not in conjunction with RT. We have included trials that are currently open as well as some that are now closed. Listed below are a few examples of potentially promising agents.

NCT00583700: This combination was studied in a randomized trial of women with breast cancer who received adjuvant radiotherapy. A statistically significant, but modest reduction in breast fibrosis was found (78).

NCT02142959: This drug is an inhibitor of NFκB (79). It was studied in a randomized trial for its protective effect in mitigating radiation-induced dermatitis in women with breast cancer who were treated with radiotherapy. Results have not been reported.

BMX-001, a redox-active Mn porphyrin, is currently in four trials. A phase I trial (NCT03608020) is a lead in trial to a randomized phase II trial of patients with multiple brain metastases, undergoing radiotherapy. A phase II trial (NCT02655601) examines whether BMX-001 protects against neurocognitive loss after temozolomide + radiotherapy for the treatment of primary glioblastoma. Two other trials test whether BMX-001 can protect mucosa and salivary glands from acute and chronic damage, respectively, after CRT for head and neck cancer (NCT02990468) or mucosal radioprotection for anal cancer (NCT03386500).

GC4419 is a Mn–macrocylic complex and a superoxide dismutase mimetic (80). This drug selectively converts superoxide anion to peroxide. The initial phase I trial of this agent was completed (NCT01921426). GC4419 was recently approved by the FDA for the prevention of mucositis patients with in head and neck cancer undergoing CRT.

As discussed above, there are several new agents that may prevent or reduce severity of both acute and chronic injury. These could be tested in clinical trials relatively easily, given that several are being or have been studied in human patients in other contexts already. As chronic radiation toxicities can develop over longer time periods, and occur in a minority of patients, designing and funding a clinical trial that shows therapeutic benefit in all-comers is challenging. One option would be to identify patients at greater risk for developing chronic injury. The previously discussed polymorphisms of PAI-1 and TGFβ or full single-nucleotide polymorphism (SNP) analyses of germline DNA could prove useful for patient selection. In addition, a stronger case for pretreatment patient selection would be made if preclinical evidence exists that mitigators are effective in high-risk populations. Earlier in this review, we discussed preclinical animal models where this hypothesis could be tested. A second trial option would be to select agents that protect against both acute and chronic consequences of radiotherapy. In this type of trial, the approval path could focus on mitigation of acute effects, but chronic tissue damage could also be studied as a secondary endpoint. Severity of acute effects was correlated with higher risk of chronic toxicity in some clinical analyses, as discussed earlier (37, 39–41), lending some justification for this approach.

There are challenges to running trials to test mitigators of pelvic radiation disease: (i) no uniform set of definitions for what constitutes pelvic radiation disease exists. This is partly because multiple organs can be affected and the involvement varies between patients. (ii) The true efficacy of commonly used interventions for pelvic radiation disease has not been established through clinical trials. This makes it difficult to define the appropriate control arm. In any trial, it will be important to develop or utilize tools, instruments, and patient-reported outcomes that quantify pelvic organ function. Using such tools, it may be possible to identify the protective value of a mitigator, even in the absence of clinically symptomatic damage.

It will be imperative to define a unified assessment tool by which clinical symptoms can be graded. The Late Effect Normal Tissue Task Force (LENT)/subjective, objective, management, analytic (SOMA) scale for late effects toxicity scoring is well-validated and one of the most widely used classification systems. The LENT/SOMA scale, however, may be not sensitive as some other scales [NCI-Common Terminology Criteria for Adverse Events (CTCAE) and Radiation Therapy Oncology Group (RTOG)/European Organization for Research and Treatment of Cancer (EORTC); refs. 81–84]. Keeping the issues discussed in mind, it is predicted that excellent randomized studies can be designed and conducted.

Technological improvements in radiotherapy have decreased the risk of long-term side effects, however, not eliminated pelvic radiation disease in all patients. It is hoped that many of the approaches discussed in this article will move into well-designed clinical trials of mitigators of pelvic radiation disease, with an ultimate goal of allowing for optimal dose to targets with a concurrent further reduction of chronic toxicity risk.

M. W. Dewhirst reports receiving commercial research grants from Biomimetix and is a coinventor for the use of BMX-001 for oncology applications (this patent application has been filed with the U.S. Patent Office). Duke University holds the patent for BMX-001, a redox-active Mn porphyrin, discussed in this paper. No potential conflicts of interest were disclosed by the other authors.

This work was supported by Duke Radiation Oncology Departmental sources and from Samsung Medical Center, which supported Dr. Huh for his sabbatical leave while working on this manuscript. For Fig. 1, the authors thank Kathleen A. Ashcraft, PhD (Department of Radiation Oncology, Duke University Medical Center) for preparing the figure. Dr. Ashcraft has agreed to be acknowledged for her support of this manuscript. For Table 1, we thank Dr. Honglaing Liu for assistance in defining the prevalence of SNPs for PAI-1 and TGFβ.