OCR Indicates Ontario Cancer Registry; SCCA, squamous cell carcinoma of the anal canal.
A. Association between radiation treatment completion status and survival. B, Association between chemoradiation treatment completion status and survival. C, Association between radiation treatment interruption status and survival.
Description of Data Sources
eTable 1. Characteristics of the study population
eTable 2. Characteristics associated with treatment completion and interruption among patients who received at least 1 dose of concurrent chemotherapy
eTable 3. Exploration of the magnitude of confounding of selected covariates
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Raphael MJ, Ko G, Booth CM, et al. Factors Associated With Chemoradiation Therapy Interruption and Noncompletion Among Patients With Squamous Cell Anal Carcinoma. JAMA Oncol. 2020;6(6):881–887. doi:10.1001/jamaoncol.2020.0809
How often do patients with anal cancer experience treatment deviations from planned curative-intent chemoradiation?
In this population-based cohort study of 1125 patients with anal cancer, treatment interruptions occurred in 262 (23%) and radiation and chemoradiation noncompletion occurred in 199 (18%) and 280 (25%), respectively. Patients who did not complete chemoradiation had a higher risk of requiring salvage surgery, cancer-specific death, and overall death.
Quality improvement efforts are needed to optimize treatment continuity and completion.
Definitive chemoradiation for anal cancer is effective but may be associated with toxic effects, and some patients may not be able to complete the planned treatment. Identifying factors associated with treatment interruption and noncompletion is important to target quality improvement efforts.
To identify rates of chemoradiation treatment interruption or noncompletion and factors associated with this among patients with anal cancer treated in routine clinical practice.
Design, Setting, and Participants
In this population-based, retrospective cohort study, the Ontario Cancer Registry was used to identify all incident cases of squamous cell anal cancer treated with curative-intent radiation from 2007 to 2015 in Ontario, Canada. Final analysis of data was performed on August 9, 2019.
Curative-intent radiation therapy.
Main Outcomes and Measures
Treatment interruption was defined as more than 7 days between fractions of radiation. Radiation completion was defined as receipt of 45 Gy or more and 25 fractions of radiation. Chemoradiation completion was defined as radiation completion and 2 doses of combination chemotherapy. Associations between patient factors and treatment interruption and noncompletion were estimated with log-binomial models. Cox proportional hazard models were used to estimate the association of treatment interruption or noncompletion with all-cause death, cancer-specific death, and the combined outcome of colostomy or death.
Overall, 1125 patients with stage I-III anal cancer were treated with curative-intent radiation. Treatment interruptions occurred in 262 (23%). Radiation and chemoradiation noncompletion occurred in 199 (18%) and 280 (25%), respectively. No associations were found to correlate with an increased risk of treatment interruption. Patients older than 70 years were less likely to complete chemoradiation (risk ratio [RR], 0.60; 95% CI, 0.52-0.70), compared with those younger than 50 years. Patients with a higher number of comorbidities were also less likely to complete chemoradiation (RR, 0.70; 95% CI, 0.51-0.95). Patients who did not complete chemoradiation had a higher risk of requiring salvage abdominoperineal resection (RR, 1.54; 95% CI, 1.03, 2.31), overall death (hazard ratio [HR], 1.54; 95% CI, 1.23-1.92), cancer-specific death (HR, 1.59; 95% CI, 1.14-2.22), and colostomy or death (HR, 1.80; 95% CI: 1.10-2.93). Treatment interruptions longer than 7 days were not associated with death.
Conclusions and Relevance
Many patients undergoing curative-intent chemoradiation for anal cancer experienced treatment interruption or noncompletion. Quality improvement initiatives to optimize treatment continuity and completion are needed.
Squamous cell carcinoma of the anal canal (SCCA) is an uncommon malignant disease accounting for less than 1% of new cancer diagnoses in the US.1 Although rare, the incidence has increased dramatically over the past 2 decades.2 Historically, radical surgery with abdominoperineal resection (APR) was the standard treatment for SCCA. Efforts to improve rates of organ preservation and local control have established combined modality chemotherapy and radiation as the standard first-line treatment for most SCCA.3,4 Compared with radiotherapy alone, the addition of chemotherapy to radiation is associated with improved locoregional control, colostomy-free survival, and death from SCCA.5,6 Several phase 3 trials have established 2 doses of infusional 5-fluorouracil and 1 or 2 doses of mitomycin concurrently with 45.0 to 50.4 Gy of irradiation in 20 to 25 fractions as the standard of care treatment regimen.5,7-9
Combined modality chemoradiation is effective but has potentially toxic effects. A considerable proportion of patients may not be able to complete the planned treatment course. For example, in the ACT-I trial,5 585 patients were randomized to 45 Gy in 20 to 25 fractions over 4 to 5 weeks with or without chemotherapy (5-fluorouracil during the first and last weeks of radiotherapy and mitomycin on day 1). In this trial, 15% of patients were reported to have at least 1 treatment interruption during chemoradiotherapy and 9% did not complete the planned total radiation dose.
The factors associated with treatment interruption and noncompletion in patients with SCCA are poorly described. Moreover, it is unknown to what extent failure to complete treatment as prescribed is associated with compromised local control and survival. Therefore, the primary objective of this study was to identify rates of, and factors associated with, chemoradiation treatment interruption or noncompletion among patients with SCCA treated in routine clinical practice. The secondary objective was to assess whether treatment interruption or noncompletion are associated with inferior clinical outcomes.
This population-based, retrospective cohort study included all patients with SCCA treated with curative-intent radiotherapy during 2007 to 2015 in the Ontario, Canada. In Ontario, radiotherapy is exclusively provided in 16 cancer centers. Each center is mandated to submit chemoradiation treatment details to Cancer Care Ontario. This study was designed, analyzed, and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.10 The study was approved by the Research Ethics Board of Queen’s University, Kingston, Canada. The need for informed consent was waived because all data used were were routinely collected administrative data.
The Ontario Cancer Registry (OCR) was used to identify all incident cases of SCCA diagnosed in the province of Ontario between January 1, 2007, and December 31, 2015. Patients were eligible for inclusion in this study if they were aged 18 years or older at the time of cancer diagnosis and received radiation treatment with curative intent, as documented by the treating radiation oncologist at the time of treatment planning (eFigure 1 in the Supplement).
A complete description of the validated, population-level databases and algorithms used to identify patients, covariates, exposures, and outcomes can be found in the eAppendix in the Supplement.
Treatment interruption was defined as more than 7 days between consecutive radiation fractions. Initially, radiation completion was defined as receipt of at least 50 Gy and 25 fractions of radiation based on the results of the ACT-II trial.8 However, we noted there was considerable regional variation in the completion of radiation across the province raising questions about the accuracy of this strict treatment definition. We therefore revised our definition of radiation completion to be at least 45 Gy and 25 fractions (the minimum acceptable radiation dose according to international clinical practice guidelines3). Chemoradiation completion was defined as radiation completion and receipt of 2 doses of 5-fluorouracil, given at least 2 weeks apart, and 1 dose of either mitomycin or cisplatin.
For all time-to-event analyses, time zero was the first dose of radiation and follow-up was until death (all-cause or cancer-specific), colostomy (colostomy-free survival), or end of follow-up. Patients were censored at the end of OHIP eligibility, maximum follow-up date (December 31, 2017), or 10 years from the first date of radiation. Cause of death information was available until December 31, 2015; therefore, patients with missing cause of death data were censored for cancer-specific survival analyses at December 31, 2015.
Descriptive frequencies are presented as proportions, counts and means (SDs), as appropriate. Pearson χ2 tests were used to test for the difference in proportions. Associations between patient factors and treatment interruption and noncompletion were estimated with log-binomial models. Cox proportional hazard models were used to estimate the association of treatment interruption and/or noncompletion with all-cause death, cancer-specific death, and the combined outcome of colostomy or death.
A priori, we selected potential confounders for adjustment in multivariable models based on knowledge of factors that might causally affect the study exposure and also affect the study outcome independent of the exposure. These factors included age, social economic status (SES), sex, stage, comorbidity-score, and HIV status. Potential confounders that produced a change of more than 1% in the association between at least 1 exposure and death were retained in the final model, except for SES, which could not be retained owing to model convergence issues. Missing radiation dose data was imputed using the median daily dose for the patient cohort (1.8 Gy). Analyses were performed using SAS statistical software (version 9.4, SAS Institute). In accordance with the policies of the International Credential Evaluation Service, to limit the risk of patient reidentification, any small cell counts are reported as fewer than 6.
Between January 01, 2007, and December 31, 2015, 1593 patients were diagnosed with SCCA in Ontario (Figure 1). Among this cohort, 1125 patients met the study inclusion criteria (eFigure 1 in the Supplement). The mean (SD) age of patients was 62 (11.9) years, 784 were women (70%), 94 were HIV positive (8%), and 967 resided in an urban setting (86%) (eTable 1 in the Supplement).The number of patients treated at each cancer center ranged from fewer than 6 to 32 patients per year (mean, 9 patients per year).
A defunctioning colostomy was required in 85 of 1125 patients (8% ) before starting or during radiation; 63 of these patients (74%) did not have the colostomy reversed after treatment completion. A permanent colostomy was required in 156 of all 1125 patients (14%), including the 63 who did not have a reversal of defunctioning colostomy and 93 who underwent salvage APR for persistent or recurrent disease.
The time from diagnosis to first day of radiation treatment was less than 12 weeks in 990 patients (88%). The mean time from diagnosis to treatment did not differ by treatment interruption: 59.6 days in those with treatment interruptions and 61.1 days in those without interruption (P = .40). The mean time from diagnosis to treatment start differed by treatment completion status: 66.6 days among patients who completed radiation and 59.4 days among those who did not (P < .01).
The median dose of radiation was 54 Gy (interquartile range [IQR], 50-59 Gy) and the median duration was 46 days. More than 25 fractions of radiation were delivered in 924 patients (82%) and more than 45 Gy of total dose was delivered in 950 patients (84%).
At least 1 dose of concurrent combination chemotherapy was administered to 1002 patients (89%): 83% (934/1125) received 5-fluorouracil with mitomycin and 68 (6%) received 5-fluorouracil with cisplatin. Two doses of concurrent chemotherapy were administered to 850 of patients (76%).
A total of 262 patients (23%) experienced treatment interruptions longer than 7 days. The proportion of patients who experienced treatment interruptions longer than 7 days varied by treatment facility: from 9% to 55%.
A total of 926 patients (82%) completed radiation and 845 (75%) completed chemoradiation. The proportion of patients who completed radiation varied by treatment facility: from 66% to 93%. The proportion of patients who completed chemoradiation varied by treatment facility: from 17% to 80%. The proportion of patients who completed chemoradiation increased over time: 55% (186/336) from 2007 to 2009, 70% (256/368) from 2010 to 2012, and 72% (303/421) from 2013 to 2015 (P for trend, <.001).
There were no factors associated with treatment interruption (Table 1). In adjusted models, patients older than 70 years were less likely to complete chemoradiation (RR, 0.60; 95% CI, 0.52-0.70), compared with those younger than 50 years. Patients with a higher comorbidity were also less likely to complete chemoradiation (RR, 0.70; 95% CI, 0.51-0.95). There was no association between patient sex or HIV status and treatment completion. Results were similar in analyses limited to those patients who received at least 1 dose of concurrent chemotherapy (eTable 2 in the Supplement).
The unadjusted 5-year overall survival (OS) was 73% (95% CI, 70%-76%) vs 61% (95% CI, 54%-68%) among patients who did and did not complete radiation treatment, respectively (P < .001). The unadjusted 5-year OS was 77% (95% CI, 73%-80%) vs 60% (95% CI, 54%-65%) among those who did and did not complete concurrent chemotherapy, respectively (P < .001) (Figure 2).
In adjusted models (Table 2), patients who did not complete radiation appeared to have a higher risk of requiring salvage APR (RR, 1.48; 95% CI, 0.94-2.34) and a higher risk of colostomy or death (HR, 1.64; 95% CI, 0.93-2.90), although these did not reach the threshold of statistical significance. Patients who did not complete radiation had higher cancer-specific death (HR, 1.90; 95% CI, 1.32-2.75) and overall death (HR, 1.47; 95% CI, 1.14-1.89). Similarly, patients who did not complete chemoradiation had higher risk of requiring salvage APR (RR, 1.54; 95% CI, 1.03-2.31), a higher risk of colostomy or death (HR, 1.80; 95% CI, 1.10-2.93), cancer-specific death (HR, 1.59; 95% CI, 1.14-2.22), and overall death (HR, 1.54; 95% CI, 1.23-1.92). Patients experiencing treatment interruptions longer than 7 days did not have worse rates of salvage APR, or survival.
Several important findings have emerged in this contemporary study of patients undergoing curative intent treatment for SCCA in routine clinical practice. First, nearly 1 in 4 patients will experience a delay during treatment longer than 1 week. Second, even with a definition of radiation completion of 45 Gy and 25 fractions, 1 in 4 patients did not complete chemoradiotherapy. Third, there was considerable variation in the proportion of patients experiencing treatment interruptions and treatment noncompletion between treatment facilities. Fourth, in adjusted models, age older than 70 years and higher comorbidity scores were the strongest patient factors associated with treatment noncompletion. These findings highlight the need for quality improvement efforts to support elderly and comorbid patients during treatment, and to ensure consistency of radiation provision across treatment facilities. In addition, it would be important to investigate the specific needs of these patient populations that are resulting in treatment protocol deviations.
Evidence across multiple cancer types supports that interruptions to radiation treatment may compromise treatment efficacy.11-16 Tumor cell repopulation, which may be accelerated by chemoradiation treatment,12,17 can occur during treatment interruptions leading to worse clinical outcomes. As a result, uninterrupted treatment without planned gaps is considered the most effective radiobiological treatment for SCCA.4 In the clinical trial setting, treatment interruptions during chemoradiation for SCCA occur in approximately 15% of patients.5,8 In the ACT-II trial, 940 patients were randomized in a 2 × 2 factorial design, 1 factor of which investigated whether mitomycin or cisplatin was the better complement to 5-fluorouracil. In this trial, 92% of patients in the mitomycin arm completed the planned 50.4 Gy radiation course, but 15% (70/472) experienced a treatment interruption during radiotherapy; 13% (59/472) of patients had interruptions of 7 days or less (median 2 days) and 2% (59/472) had interruptions longer than 7 days (median 12 days). In this study of patients treated in routine clinical practice, 262 (23%) and 203 (18%) patients experienced treatment interruptions longer than 7 and longer than 10 days, respectively. Given the well-described differences between patients enrolled to clinical trials and those treated in routine clinical practice,18 the higher rate of treatment interruption is expected. The information is important to properly counsel patients about expected course of treatment and to guide quality improvement initiatives to minimize treatment interruptions.
In this study, 926 patients (82%) completed radiation and 845 (75%) completed radiation and received 2 doses of combination chemotherapy. Patients older than 70 years and those with higher comorbidity scores were less likely to complete radiation and receive 2 doses of concurrent chemotherapy. We are unable to determine if this was related to toxic effects of treatment or if clinicians planned to deliver lower doses and omit chemotherapy for these patients; because elderly and more comorbid patients did not have higher rates of treatment interruption, we hypothesize the latter is the more likely reason. Although there are conflicting reports as to how older adults tolerate chemoradiation treatment for SCCA,19-23 and some studies have suggested omission of chemotherapy or reduction in radiation dose/volume might be appropriate,24,25 guidelines still recommended that older adults be treated similarly to their younger counterparts.3,4 Therefore, the results of this study highlight the need for audits and quality improvement efforts to ensure that older adults are receiving guideline-concordant treatment. In particular, the results of this study emphasize the importance of treatment completion. The unadjusted 5-year overall survival was 72% vs 61% among patients who did and did not complete radiation treatment, respectively, whereas the unadjusted survival among those who did and did not have treatment interruptions did not differ. Thus, our results show that there were 12 more patient deaths per 100 patients who stopped treatment than there were per 100 patients who completed treatment.
The results of this study are important to consider in the context of its limitations. Imbalances in unmeasured and unknown prognostic factors also related to treatment decisions may have confounded our results. Although we adjusted for some potential confounders, we lacked information on patient performance status, tumor size, and nodal status. Although we did adjust results for patient comorbidity, this is an imperfect correlate of performance status. Residual confounding owing to possible unmeasured time-dependent confounders (ie, changes in cancer severity/overall health from the time of cancer diagnosis to treatment start) should also be considered as a study limitation. Although we know the intent of treatment at the time of radiation planning, the original planned radiation dose and the specific volumes/fields prescribed are not known. There was variability in rates of treatment noncompletion across centers, raising the possibility that each center had a different definition of what a full treatment dose should be. However, clinical guidelines recommend doses of at least 45 Gy,3,4 even for stage 1 disease, thus our definition should have a high sensitivity for correctly classifying planned treatment. Similarly, some treatment protocols call for 2 doses of mitomycin (during week 1 and week 5) and others call for only 1. We cannot ascertain whether some patients were originally planned for 2 doses of mitomycin and only received 1. Thus, our definition of complete chemoradiation may overestimate the rate of planned chemoradiation completion. Finally, we did not have information on intensity modulated radiation therapy (IMRT) usage. Intensity modulated radiation therapy has been associated with reduced overall treatment time and toxic effects,26,27 which may result in fewer treatment interruptions and higher chance of planned treatment completion. In this study, we identified improvements in the rate of treatment completion and less treatment interruptions in more recent years. This may reflect increased adoption of more sophisticated radiation techniques, such as IMRT.
In this large, contemporary population-based study of patients undergoing curative intent chemoradiation for SCCA, we identified that many patients will experience treatment interruption and will not complete the planned course. Quality improvement initiatives to optimize treatment, treatment continuity, and completion are needed. Particular attention is needed to ensure that older and comorbid adults have the supports they need to successfully complete treatment.
Corresponding Author: Sunil V. Patel, MD, MSc, Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, 10 Stuart St, Kingston, ON K7L 3N6, Canada (Sunil.Patel@kingstonhsc.ca).
Accepted for Publication: March 2, 2020.
Published Online: April 23, 2020. doi:10.1001/jamaoncol.2020.0809
Author Contributions: Drs Patel and Li had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Raphael, Ko, Booth, Brogly, Kalyvas, Hanna, Patel.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Raphael, Booth, Kalyvas, Hanna, Patel.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Raphael, Li, Hanna.
Obtained funding: Hanna, Patel.
Administrative, technical, or material support: Ko, Brogly, Kalyvas, Hanna, Patel.
Supervision: Kalyvas, Hanna, Patel.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Christopher Booth is supported as the Canada Research Chair in Population Cancer Care. Dr Timothy Hanna holds a research chair provided by the Ontario Institute for Cancer Research through funding provided by the Government of Ontario (#1A-035). Funding for this study was provided by the Queen’s University Department of Surgery Research Fund.
Role of the Funder/Sponsor: The Queen’s University Department of Surgery Research Fund had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: The authors thank Katherine Lajkosz, MSc, International Credential Evaluation Service, Queen’s, who contributed to developing the patient cohort and statistical analyses.
Create a personal account or sign in to: