Powered by the Evidence-based Practice Centers
Evidence Reports All of EHC
Evidence Reports All of EHC

SHARE:

FacebookTwitterFacebookPrintShare

Treatment of Stage I-III Squamous Cell Anal Cancer

Systematic Review Aug 20, 2024
Download files for this report here.

  • For the initial treatment of stages I–III squamous cell carcinoma of the anus (SCCA), compared with radiation therapy (RT) alone, concurrent doublet chemoradiation (CRT) with 5-fluorouracil (5FU) plus mitomycin (MMC) likely results in greater disease-free survival and lower locoregional failure rate, may result in greater colostomy-free survival, likely increases overall acute harms, may increase hematologic toxicity, and may have no difference in late harms.
  • Compared with concurrent CRT with 5FU, CRT with 5FU plus MMC may result in greater disease-free and colostomy-free survival and lower locoregional failure rate but difference in harms remains uncertain.
  • Compared with CRT with 5FU plus MMC, CRT with 5FU plus cisplatin likely does not increase overall or progression-free survival and complete response rates, likely does not decrease distant metastasis rate; and may not increase disease-free survival, nor lower locoregional failure rate (up to 5 years of followup), likely decreases the risk of hematologic toxicity, but may not differ in other acute or late harms. Difference in colostomy-free survival remains uncertain.
  • Compared with doublet CRT with capecitabine plus MMC, triplet CRT with paclitaxel plus capecitabine plus MMC may increase complete response, overall, disease-free, and colostomy-free survival, and overall acute harms, but may not differ in acute neutropenia, dermatologic, gastrointestinal, or genitourinary toxicity.
  • Evidence remains uncertain for other comparisons including local excision versus CRT for stage I cancer; capecitabine versus 5FU; different modalities, doses, volumes, and fractionation schema for RT; dose de-escalation or escalation in CRT; immunotherapy; posttreatment surveillance strategies; and patient-reported outcomes such as bowel, urinary, and sexual function, pain, and quality of life.
  • Most of the studies are assessed with high risk of bias. Observational studies should apply advanced methods such as target trial emulation for making causal inferences.
  • Patients with immunocompromised status, older age, or minoritized racial/ethnic identities are underrepresented in the available body of evidence.

Summary of
Findings
Key
Questions
Overview
Clinical and Policy
Implications
Caveats, Applicability,
and Limitations
 

We included 33 articles from eight RCTs (six with low to moderate risk of bias [RoB] and two with high RoB) and 20 NRSIs (serious to critical RoB). Compared with RT alone, CRT with 5FU plus MMC improved several effectiveness outcomes (moderate to low SOE) but increased overall acute harms (moderate SOE) and resulted in no difference in late harms (low SOE). Compared with CRT with 5FU, CRT with 5FU plus MMC improved several effectiveness outcomes (low SOE), but evidence was insufficient to compare harms. Compared with CRT with 5FU plus MMC, CRT with 5FU plus cisplatin did not improve several effectiveness outcomes and late harms (moderate to low SOE), lowered acute hematologic toxicity (moderate SOE), and evidence was insufficient and conflicting for colostomy-free survival. Compared with CRT with capecitabine plus MMC, CRT with paclitaxel plus capecitabine plus MMC showed greater overall, disease-free, and colostomy-free survival, and overall acute harms (low SOE). Evidence was insufficient for remaining comparisons including local excision for early-stage cancer, capecitabine versus 5FU, intensity modulated versus three dimensional conformal RT, proton versus photon beam, external beam RT versus brachytherapy boost, different RT doses, volumes, and fractionation schema, induction therapy, maintenance therapy, one versus two cycles of MMC, RT boost versus no boost, immunotherapy, and posttreatment surveillance. Evidence was insufficient for patient-reported outcomes such as bowel, urinary, and sexual function.

Summary of key findings and strength of evidence

ComparisonOutcomeStudy Number, Design, and Participants (n)Summary of Individual Study Reported FindingsSOE
CRT with 5FU plus MMC vs. RT alone
new tab
Overall survival (up to 5 yrs)2 RCTs; n=695No significant difference; precise estimates.Moderate 
moderate SOE
 
Disease-specific mortality (up to 5 yrs)1 RCT; n=585CRT favored over RT alone; rate, 28% vs. 39%, RR of 0.71 (95% CI, 0.53-0.95, median follow up: 42 months and similar in both arms)Moderate moderate SOE
 
CR (6 weeks post treatment)2 RCTs; n=695Both RCTs favor CRT over RT alone, rates 80% vs. 54% (p<0.05, n=110); 39% vs.30% (p<0.05, n=585)Low low SOE
 
LR failure rate (up to 5 yrs)2 RCTs; n=695Both RCTs favor CRT over RT alone, with 5-year rate 32% vs. 50% (p= 0.02, n=110) and 3-year rate 39% vs. 61% (RR, 0.54; 95% CI, 0.42 - 0.69; n=585)Moderate moderate SOE
 
Colostomy-free survival (up to 5 yrs)1 RCT; n=110CRT favored over RT alone; estimated improvement by 32% at 5yrs (p=0.002); estimates for each arm not reportedLow low SOE
 
Overall acute harms1 RCT; n=585Significantly greater in CRT (48%) vs. RT alone (38.6%).Moderate moderate SOE
 
Acute hematologic toxicity2 RCTs; n=695Significantly greater in CRT vs. RT alone, frequency: 1 vs.0 grade 4 events (n=110); or 20 vs.0 overall events (n=585).Low low SOE
 
Acute derm and GI toxicity2 RCTs; n=695No significant difference, imprecise estimates.Low low SOE
 
Acute GU toxicity1 RCT; n=585No significant difference, imprecise estimatesLow low SOE
 
Overall late harms1 RCT; n=585No significant difference, imprecise estimates.Low low SOE
 
Late derm toxicity2 RCTs; n=695No significant difference, imprecise estimates.Low low SOE
 
Late GI and GU toxicity1 RCT; n=585No significant difference, imprecise estimates.Low low SOE
 
CRT with 5FU plus MMC vs. CRT with 5FU
new tab
Overall survival (at 4 yrs)1 RCTs; n=310No significant difference, precise estimates.Low 
low SOE
 
CR (4-6 weeks post treatment)1 RCTs; n=310No significant difference, precise estimates.Low low SOE
 
DFS (at 4 yrs)1 RCTs; n=310Favors 5FU + MMC over 5FU alone, 73% vs. 51% (p<0.001).Low low SOE
 
LR failure rate (at 4 yrs)1 RCTs; n=310Favors 5FU + MMC over 5FU alone, 16% vs 34% (p<0.001).Low low SOE
 
Colostomy-free survival (at 4 yrs)1 RCTs; n=310Favors 5FU + MMC over 5FU alone, 71% vs. 59% (p = 0.014).Low low SOE
 
CRT with 5FU plus MMC vs. CRT with 5FU plus cisplatin
new tab
Overall survival (up to 5 yrs)2 RCT; n=1622No significant difference, precise estimates.Moderate 
moderate SOE
 
DM (up to 5 yrs)1 RCT; n=682No significant difference, precise estimates.Moderate moderate SOE
 
LR failure and DFS (up to 5 yrs)1 RCT; n=682No significant difference, imprecise estimates.Low low SOE
 
PFS, CR1 RCT; n=940No significant difference, precise estimates.Moderate moderate SOE
 
Overall acute harms2 RCT; n=1622No significant difference, precise estimates.Moderate moderate SOE
 
Acute hematologic toxicity2 RCT; n=1622Significantly greater with MMC vs. cisplatin, with grade 3+ toxicity frequency of 61% vs. 42% (p<0.001, n=682) and 26% vs. 16% (p<0.001, n=940).Moderate moderate SOE
 
Acute derm, GI, and GU toxicity2 RCT; n=1622No significant difference, imprecise estimates.Low low SOE
 
Late harms1 RCT; n=682No significant difference, imprecise estimates.Low low SOE
 
CRT with capecitabine plus MMC plus paclitaxel vs. CRT with capecitabine plus MMCOverall survival (up to 3 yrs)1 RCT; n=144Significantly greater in the paclitaxel arm vs no paclitaxel arm; 95.5% vs 80% (p<0.001).Low 
low SOE
 
Disease-free survival (up to 3 yrs)1 RCT; n=144Significantly greater in the paclitaxel arm vs no paclitaxel arm; 87.1% vs 64.4% (p=0.001).Low low SOE
 
Colostomy-free survival (up to 3 yrs)1 RCT; n=144Significantly greater in the paclitaxel arm vs no paclitaxel arm; 83.2% vs 67.5% (p=0.029).Low low SOE
 
CR (at 26 weeks posttreatment)1 RCT; n=144Significantly greater in the paclitaxel arm vs no paclitaxel arm; 88.9% vs 75% (p=0.049).Low low SOE
 
Overall acute harms1 RCT; n=144Significantly greater in the paclitaxel arm vs no paclitaxel arm; 56.9% vs 26.4% (p<0.001).Low low SOE
 
Neutropenia (grade 3 or 4)1 RCT; n=144No significant difference, imprecise estimates.Low low SOE
 
Acute derm, GI, GU harms1 RCT; n=144No significant difference, imprecise estimates.Low low SOE
 
     
Note: Remaining comparisons not mentioned in the table received an insufficient grade. All harms are grade 3 or 4 unless specified otherwise. Abbreviations: LR- locoregional, CR- complete response, OS-overall survival; DFS- disease-free survival; PFS- progression-free survival; DM- distant metastasis; 5FU- 5 fluorouracil, MMC- mitomycin C; RT- radiation therapy; derm-dermatologic; GI-gastrointestinal; GU-genitourinary; RCT- randomized controlled trial; yrs- years; vs.- versus; RR- relative risk.
 
Where links are available within the Report Snapshot tables, clicking the link will take you to the PubMed listing for the studies available within PubMed. Not all studies in all findings are available in PubMed.

The Evidence-based Practice Center drafted the following key questions (KQs) with input from public comments, key informants, and a panel of technical experts.

KQ 1: What are the effectiveness and harms of different modalities of initial treatment for stages I-III squamous cell anal cancer?

KQ 2: What are the effectiveness and harms of different modalities of radiation therapy for initial treatment of stages I-III squamous cell anal cancer?

KQ 3: What are the effectiveness and harms of different radiation therapy doses, volumes, and fractionation schema for initial treatment of stage I-III squamous cell anal cancer?

KQ 4: What are the effectiveness and harms of different combinations of chemotherapy and radiation therapy, and dose de-escalation or dose escalation for initial treatment of stages I-III squamous cell anal cancer?

KQ 5: What are the effectiveness and harms of immunotherapy for initial treatment of stages I-III squamous cell anal cancer?

KQ 6: What are the effectiveness and harms of different frequencies and modalities for posttreatment surveillance strategies after initial treatment of stages I-III squamous cell anal cancer?

For all KQs, do the outcomes differ by patient characteristics such as age, sex, immunocompromised status, or other characteristics associated with health inequities (such as race/ethnicity)?

Overview

Using exclusively comparative effectiveness studies, our findings aimed to help identify optimal strategies for initial treatment of nonmetastatic SCCA including local excision (LE) in early-stage disease; the optimal chemotherapy regimen; ideal radiation therapy (RT) technique and dose fractionation scheme; potential roles for immunotherapy; and effective posttreatment surveillance. The review covered a diverse set of interventions intended to maximize tumor response while limiting recurrence and treatment-related toxicities. However, due to challenges with high risk of bias (RoB) precluding the ability to draw causal inferences, we relied on findings from only six RCTs. We identified moderate- to low- strength evidence to support optimal chemoradiation (CRT) regimens. We found little to no information on various RT dosing regimens or modalities, or posttreatment surveillance. Furthermore, many studies were conducted before the widespread implementation of intensity modulated RT (IMRT), making the findings less applicable to current practice.

We found low- and moderate-strength evidence in two RCTs indicating that compared with CRT, RT alone resulted in greater locoregional failure rate and disease-specific mortality rates and lower CR and CFS rates while also incurring lower acute hematologic and overall harms. But low-strength evidence also showed that no other acute or late toxicities differed between CRT and RT alone. Moderate- to low- strength evidence from two RCTs showed that compared with CRT with 5FU and MMC, CRT with 5FU and cisplatin was not superior for oncological outcomes but resulted in lower hematologic toxicity. We found low-strength evidence from one RCT indicating that adding paclitaxel as a third cytotoxic agent to doublet CRT with capecitabine plus MMC may improve several effectiveness outcomes and increase treatment related acute toxicity. For all other interventions and outcomes, we found the evidence insufficient. This does not mean that none of the individual interventions described are potentially useful for patients, practitioners, and the healthcare system. Rather, it means that current available evidence cannot yet provide clear answers about which, if any, interventions offer consistently greater benefits on a relative scale. At present, the uncertainty of the evidence is too high for us to draw many conclusions. Further, when studies with inadequate power, no power calculations, or other serious methodological limitations report no significant differences between interventions, it is inappropriate to conclude non-inferiority of any interventions.

Implications for Clinical Practice and Policy

The lack of sufficient evidence to support widespread dissemination of several interventions analyzed in this review leaves patients, their families, and practitioners without clear answers. Owing to under-representation in the available body of evidence, decision making for patients with immunocompromised status, older age, and minoritized racial/ethnic identities, who face disproportionately worse treatment outcomes, remains challenging. When deciding on treatment approaches, institutions and providers will continue to depend on limited, best available evidence along with subjective observations and clinical gestalt till more evidence becomes available; while taking individual patient characteristics and preferences into consideration. Future RCTs should consider implementing policies to intentionally increase representation of historically underrepresented patient subgroups.

Key Findings in Relation to Clinical Dilemmas

Systemic therapy

CRT is the standard of care for nonmetastatic SCCA; but treatment approaches vary widely, as do reported outcomes, making between-study comparisons difficult. Treatment-related morbidity is considerable, yet poorly captured, with inconsistent definitions delineating acute versus late toxicities and infrequent study of long-term toxicities. In addition, we found little data regarding optimal posttreatment surveillance and a complete lack of published data describing the current role of immunotherapy. Despite these gaps in knowledge, we sought to critically appraise the available literature to help address specific questions related to individualized treatment approaches. Though various RCTs and NSRIs have established concurrent doublet CRT as the primary treatment, the optimal regimen has yet to be established. Internationally, current recommendations support the use of concurrent doublet CRT with fluoropyrimidine (either infusional 5-FU or its oral prodrug capecitabine) along with MMC or cisplatin.

Our review confirmed that overall, doublet CRT with 5-FU and MMC resulted in better effectiveness outcomes than singlet CRT using 5FU or RT alone. Of note, in the three RCTs supporting these comparisons, the outcomes OS and CR rates should be interpreted cautiously. In these three RCTs, CR was assessed between four to six weeks posttreatment which might be premature, leading to fewer events (low power) and imprecision in estimates. A post hoc study of the ACT II trial, which was conducted several years after these RCTs, suggests that the optimal time to assess may be 26 weeks after CRT initiation. Additionally, these three RCTs were conducted in the era of older, outdated RT methods which carried a much greater risk of toxicity. Therefore, the null findings for OS might be explained by death from other causes such as treatment-related toxicity. Cisplatin, when compared with MMC in a doublet chemotherapy regimen with 5FU, resulted in lower acute hematologic toxicity but no significant difference in overall acute harms and several effectiveness outcomes. Despite this, MMC has historically been favored over cisplatin based mainly on long-term results of the RTOG 98-11 trial, a preference that merits a more nuanced discussion. The long-term followup of RTOG 98-11 was a retrospective post hoc analysis of the same 5-year outcomes analyzed in the parent trial report, these analyses were not prespecified in the trial protocol, and it had a high RoB for reasons such as long-term outcome ascertainment was inadequately described and likely retrospective and it lacked a thorough competing risks analysis and adjustment for multiple comparisons testing. In addition, the parent study and the long-term followup study showed conflicting results, with 5-year colostomy-free survival improving significantly with MMC in the parent trial but not in the long-term followup study and vice versa for 5-year overall and disease-free survival. Moreover, in RTOG 9811, only the cisplatin arm had an additional induction phase delivered before concurrent CRT, which could have played a dominant role on the outcomes in this study and limits the ability to draw inferences for an exclusively concurrent CRT regimen. The subsequent ACT-II trial was deemed null; it failed to demonstrate an oncologic benefit for cisplatin over MMC, with similar toxicity profiles. Thus, the difference between MMC and cisplatin appears to be driven by MMC's worse hematologic toxicity profile. Upon detailed examination, we feel these oncologic outcomes should be questioned at best. Contextual factors might also influence clinical decision making (for example, the ease of administering MMC vs cisplatin). Subsequent cisplatin-based treatment options for progressive disease may also influence initial treatment choice. One RCT reported that the addition of paclitaxel to CRT with capecitabine plus MMC may increase treatment efficacy and toxicity. Following the InterAAct trial, paclitaxel is used to treat metastatic SCCA. However, it is not specifically recommended in guidelines for stages I-III SCCA (our population of interest). Of note, the InterAAct trial included 10% population with nonmetastatic SCCA; furthermore, there was a case series from Wisconsin, presented as a poster at a scientific meeting which reported that in 9 patients with localized anal cancer who were ineligible for 5FU + MMC CRT regimen, a carboplatin + paclitaxel CRT regimen resulted in 100% complete clinical response.

Another comparison of interest is capecitabine (an orally administered prodrug of 5FU) versus 5FU (administered as continuous intravenous infusion for 4 days in weeks 1 and 5). Capecitabine might be more convenient, pragmatic, and a cheaper alternative to infusional 5FU. Our review did not find sufficient direct evidence comparing the effectiveness and harms of these two drugs. However, some evidence in advanced gastric and colorectal cancers suggests that capecitabine is noninferior to infusional 5FU; and international guidelines recommend either capecitabine or 5FU in the treatment of nonmetastatic SCCA11. Multiple ongoing clinical trials are assessing immunotherapy as a component of initial treatment strategies (for example, NCT04230759, NCT03233711; Appendix D), but without reportable data yet.

For maintenance chemotherapy, compared with no maintenance chemotherapy, we found insufficient evidence for its impact on effectiveness outcomes. In the ACT II RCT, which evaluated maintenance chemotherapy versus none, only 44 percent of the patients randomized to the maintenance chemotherapy arm completed the regimen, reportedly due to toxicity events or patient preferences. The high rate of differential attrition in the ACT II trial induces a very high RoB in its evaluation of the impact of maintenance chemotherapy on the RCT endpoints. However, given the poor feasibility of completing maintenance chemotherapy due to greater toxicity in ACT-II trial as well as the lack of evidence of any additional benefits of maintenance chemotherapy, international guidelines recommend against providing maintenance chemotherapy for the initial treatment of stages I-III SCCA.

Therefore, despite the vast amount of low to moderate quality data available, no choice of doublet cytotoxic chemotherapy selection has clearly emerged as superior in terms of maximizing oncologic outcomes while minimizing treatment-related toxicities. As clinicians counsel patients on their individualized treatment approach, consideration must be given to co-morbid conditions, compliance, cost, drug availability, and implications for quality of life.

Radiation therapy

Advances in RT have depended on the fundamental concept that targeted delivery to diseased tissue maximizes response while limiting morbidity to nearby organs at risk. This thought process has led to increased use of IMRT. Despite these potential benefits, our review found insufficient evidence to compare the effectiveness and harms of IMRT and 3DCRT. The Phase II, single arm RTOG 0529 study was among the first to report clinically acceptable effectiveness and favorable toxicity outcomes in support of the feasibility of IMRT. However, due to methodological limitations, including the lack of a concurrent control arm or a historical control group by applying requisite statistical weighting/matching techniques, this study was not included in our analyses. Notably, the contouring protocols developed in the RTOG 0529 have been frequently cited in the literature in this field. In one pilot RCT (n=20), compared with 3DCRT arm (n=10), patients in the IMRT arm (n=10) were able to receive a lesser radiation dose to bowel, bladder, and bone marrow, and had a lower frequency of gastrointestinal toxicity. Based on this pilot RCT and the NRSIs assessed in this review, some clinical experts recommend the use of IMRT over 3DCRT. Further efforts to reduce radiation doses to nearby organ systems led to the comparison of IMPT (proton therapy) and traditional IMRT (using photons) after a feasibility study showed similar toxicity rates of IMPT to historical controls in the RTOG 0529 trial. In line with the prior feasibility study, the one NSRI in this review comparing IMPT to IMRT found no differences in acute or late toxicity outcomes.

Other modifications to radiation therapy delivery have been tested. In Europe, conventional RT practices include split-course therapy, consisting of an initial total dose of 45 Gy to the pelvis, followed by a boost dose of 15-20 Gy to the anal canal (after a 6- to 8-week gap) delivered via either EBRT or brachytherapy. A pivotal principle of radiation therapy lies in the delivery of a high total fractionated dose without prolongation of overall treatment time, since extended treatment time has a detrimental impact on local control. Evidence was insufficient to compare EBRT versus brachytherapy, although brachytherapy had a higher incidence of late ulcers/ radionecrosis in one NSRI. Similarly, no differences were observed in survival outcomes or quality of life when comparing standard-dose boost (15 Gy) versus high-dose boost (20-25 Gy). Finally, locoregional control or overall survival did not differ between patients who received a boost versus those that did not. In addition to boost therapy, varying dose fractionation schemes were assessed regarding whether de-escalation was appropriate in achieving similar survival rates. Two NRSIs found that total radiation doses <48.60 Gy, 50.40 Gy delivered over > 42 days, and ≤4.72 fractions per week (a surrogate for prolonged overall treatment time), were associated with lower overall survival. However, we found that the NRSIs had a very high RoB making it impossible to draw causal inferences.

Thus, current evidence is insufficient for assessing methods of radiation delivery, presence or absence of a boost, or fractionation schema. Furthermore, evidence concerning patient-reported outcomes is insufficient. Patient-reported outcomes such as persistently poor quality of life, pain, and worsening bowel and sexual function are potential adverse effects of chemoradiation. One RCT included in this review evaluated quality of life, but outcome data were missing in about two-thirds of the patients included in the RCT leading to a very high RoB and insufficient evidence. None of the included studies evaluated sexual function as an outcome of interest.

Posttreatment surveillance

Alongside improving CRT strategies, optimizing a posttreatment surveillance schedule bridges the gap between therapeutic innovation and health care implementation. Timely detection of local recurrence or distant metastases allows for appropriate next steps in management, such as salvage treatment. Risk stratifying tumor characteristics and determining posttreatment relapse patterns could further improve already favorable long-term outcomes. One NSRI included in this review found that 89 percent of local recurrences occurred by year 2, with the majority being found due to symptoms of anal pain, bleeding, and persistent ulceration, while the remaining were found by a 3-month posttreatment PET scan. Although the evidence is very limited, this suggests that strict adherence to surveillance regimens in the first two years is paramount, with the potential for a reduction in surveillance frequency thereafter.

This review does not provide cost information.

Strengths and Limitations

Strengths of the review process

We adopted a review scope intended to limit bias on interventions and posttreatment surveillance for nonmetastatic SCCA. We included studies with an active comparator to determine direct effects of the described interventions. While this constraint may have led to studies being overlooked, it allowed a very high-level assessment of the state of the science in this domain while attempting to maintain data quality. Because ours was a targeted systematic review, we took a rigorous approach to assessing for RoB and based our decisions on the highly variable studies included in this review as well as the complexity of care in this disease process. Studies not rising to the needed level of scientific rigor were determined to be at higher RoB, and, although we did include potentially valuable evidence from higher-RoB studies in our review, caution must be used in applying these conclusions to clinical practice.

Limitations of the review process

We excluded studies with fewer than 15 participants per study arm as well as single arm studies, preventing us from evaluating several interventions of interest; but such studies had very high RoB, making it impossible to draw causal inferences and inclusion of such studies would not have changed the conclusions of this review. We only reviewed publications available in English language in the four large databases mentioned before. Therefore, we may have missed relevant publications published in other languages or available only in other databases. We started our search from January 2000 and relied on published systematic reviews and reference lists of included studies to cover literature published before this date. Based on feedback from peer reviewers, key informants, technical experts, and content experts in our team, a federal register notice, and public comments, we did not identify any additional relevant studies which we may have missed; nonetheless, the potential for missing older relevant studies remains. We based our SOE assessment on statistical rather than clinical significance because established thresholds of clinically meaningful effects for individual outcomes do not exist. This would likely not change the conclusions of this review. We prioritized clinical outcomes based on feedback from key informants, technical experts, partners, and content experts on our team; other process outcomes such as convenience and costs might influence clinical decision making.

Limitations of the evidence base

We found no more than four studies per unique intervention-outcome comparison and most of the studies had a high RoB. Large prospective research studies are difficult to conduct in developed countries where nonmetastatic SCCA is a rare disease. Clinical experts can make treatment recommendations based on clinical experience, pragmatism, and extrapolation of findings from other cancers; however, from a systematic review perspective, we can only critique available evidence. RCTs were limited by lack of consistency in outcome definitions and measurement between trials, inadequately addressing attrition, multiple comparisons testing, and competing risks over a longer followup period, and inadequate power to compare acute and late harms. One study attempted to establish a core outcome set for SCCA research with the help of a Delphi study including healthcare professionals and patients; however, its uptake in the research community is unclear. Common limitations of the NRSIs included selection bias, poorly defined interventions and comparators, lack of power calculations, and inadequately addressing missing data, competing risks, and potential confounding. Our conclusions largely reflect weaknesses of the evidence base. Patients with immunocompromised status, older age, and minoritized racial/ethnic identities were underrepresented in research, making applicability of findings to these subgroups challenging.

Implications for Future Research

A search on October 31, 2023, in the ClinicalTrials.gov registry identified eight potentially relevant records and one record by hand search (Appendix D), one trial comparing photon radiotherapy (IMRT, VMAT, helical tomotherapy) to IMPT, several studies looking at the role of immunotherapy (nivolumab, sintilimab, and durvalumab) + CRT versus CRT alone, and, finally, one trial examining the role of circulating tumor DNA in followup.

This review identified major evidence gaps for initial treatment and post treatment surveillance of nonmetastatic SCCA which merit the focus of future research, such as:

  • Replication study to confirm whether addition of paclitaxel as a third cytotoxic agent to the standard doublet CRT regimen increases treatment effectiveness.
  • Long-term comparative effectiveness of CRT strategies for non-metastatic SCCA in terms of oncologic outcomes, toxicities, and patient-reported outcomes using advanced causal inference methods to analyze meticulously curated real-world data.
  • Comparative effectiveness of LE versus CRT in the treatment of stage I SCCA.
  • Comparative effectiveness, harms, and costs of different RT modalities, optimal dose and fractionation schema for RT, and optimal dose for chemotherapy agents.
  • Well-designed NRSIs to evaluate applicability of RCT findings in real world settings such as rural or low volume cancer care centers and evaluate long-term outcomes and harms such as sexual and bowel dysfunction.
  • Immunotherapy for the initial treatment of nonmetastatic SCCA, noting that ongoing RCTs (Appendix D) are attempting to address this issue.
  • Enhancing the data quality of large databases, with details for each component of CRT and toxicity data to complete more robust and methodologically sound comparisons.
  • Focusing on underrepresented populations (older age, immunocompromised status, and racial/ethnic minorities), because we found only scarce, low-quality data.
  • Evaluating patient-reported outcomes such as bowel function (fecal incontinence, urgency, proctitis, anal stenosis, anal sphincter control), bladder function, sexual function (dyspareunia, vaginal stenosis, penile impotence), pain, and quality of life, as well as studying the social impact on the patient’s life.
  • Evaluating best approaches for mitigating harms and optimizing quality of life posttreatment (e.g., pelvic muscle training, preventing anal and vaginal stenosis).

Some evidence gaps are under investigation in ongoing trials (Appendix D). One example is the PLATO (PersonaLising Anal cancer radioTherapy dOse protocol) which includes three anal cancer trials (ACT) known as ACT3, ACT4 and ACT5. These trials will evaluate RT dose escalation and de-escalation strategies in specific clinical scenarios for personalizing treatment strategies. In a prespecified pilot phase analysis of ACT5 RCT, comparing standard dose vs dose-escalated concurrent CRT in advanced nonmetastatic SCCA, investigators reported that RT dose intensification appeared to be safe with acceptable compliance, acute toxicity, and patient-reported outcomes (quality of life, pain, and bowel toxicity) at 6 months. However, these findings were presented at a scientific conference and not yet published as a full-length report; therefore, we did not formally include them in our analyses; although they appear promising.

Conclusions

Concurrent CRT with 5FU plus MMC is more effective but has greater acute hematologic toxicity than RT alone or CRT with 5FU, for the initial treatment of stages I-III SCCA. The addition of paclitaxel to doublet CRT with capecitabine plus MMC may increase treatment efficacy and toxicity. Evidence is insufficient for optimal posttreatment surveillance strategies, quality of life, and other patient-reported outcomes. Future RCTs should be more inclusive of historically underrepresented patient subgroups. Future real world evidence generation must prioritize methodologic rigor and apply appropriate methods such as target trial emulation for drawing causal inferences.

Objectives. To evaluate the comparative effectiveness and harms of initial treatment and posttreatment surveillance strategies for stages I–III squamous cell anal cancer.

Data sources. MEDLINE®, Embase®, Cochrane Register of Controlled Trials, and ClinicalTrials.gov from January 2000 through March 2024; reference lists of systematic reviews and included studies; and a Federal Register notice.

Review methods. Using predefined criteria and dual review, we selected randomized controlled trials (RCTs) and nonrandomized studies of interventions (NRSIs) comparing strategies for chemotherapy, radiation therapy (RT), and surgery; modalities, doses, volumes, and fractionation schema for RT; dose de-escalation or escalation in chemoradiation (CRT); immunotherapy; and posttreatment surveillance. We evaluated risk of bias (RoB) using the RoB2 tool for RCTs and the ROBINS-I tool for NRSIs and strength of evidence (SOE) using Agency for Healthcare Research and Quality Evidence-based Practice Center Program methods for prespecified outcomes (PROSPERO registration number CRD42023456886).

Results. We included 33 articles from 8 RCTs (6 with low to moderate RoB and 2 with high RoB) and 20 NRSIs (all with serious to critical RoB). Compared with RT alone, doublet CRT with 5-fluorouracil (5FU) plus mitomycin C (MMC) showed lower locoregional failure rate (LRF) and greater disease-specific and colostomy-free survival (CFS) (moderate to low SOE), greater hematologic toxicity (low SOE), greater overall acute harms (moderate SOE), and no difference in late harms (low SOE). Doublet CRT with 5FU plus MMC showed lower LRF (low SOE) and greater CFS and disease-free survival (DFS) (low SOE) than CRT with 5FU, and evidence was insufficient to compare harms. Compared with CRT with 5FU plus MMC, CRT with 5FU plus cisplatin did not improve several effectiveness outcomes up to 5 years, or overall acute or late harms (moderate to low SOE), showed lower hematologic toxicity (moderate SOE), and had conflicting, insufficient evidence for CFS. Triplet CRT with paclitaxel plus capecitabine plus MMC showed greater CFS, DFS, overall survival, and overall acute harms than doublet CRT with capecitabine plus MMC (low SOE). Remaining comparisons had insufficient evidence. Patients with older age, immunocompromised status, or minoritized racial/ethnic identities were underrepresented in included studies.

Conclusions. Doublet CRT is likely more effective but may have greater hematologic toxicity compared with RT alone or CRT with 5FU. Adding paclitaxel to doublet CRT may increase treatment efficacy and toxicity. Evidence is insufficient for optimal posttreatment surveillance strategies, quality of life, and other patient-reported outcomes. Future RCTs should increase inclusion of historically underrepresented patients, and future real-world evidence generation must prioritize methodological rigor.

Troester A, Parikh R, Southwell B, et al. Treatment of stages I-III squamous cell anal cancer: a comparative effectiveness systematic review. J Natl Cancer Inst. 2024 Aug 20:djae195. doi: 10.1093/jnci/djae195. PMID: 39163501.

Parikh RR, Troester A, Southwell B, Ester E, Sultan S, Claussen AM, Greeno E, Arsoniadis E, Church TR, Wilt TJ, Goffredo P, Butler M. Treatment of Stages I–III Squamous Cell Anal Cancer: A Systematic Review. Comparative Effectiveness Review No. 273. (Prepared by the Minnesota Evidence-based Practice Center under Contract No. 75Q80120D00008.) AHRQ Publication No. 24-EHC026. PCORI® Publication No. 2024-SR-03. Rockville, MD: Agency for Healthcare Research and Quality. August 2024. DOI: https://doi.org/10.23970/AHRQEPCCER273. Posted final reports are located on the Effective Health Care Program search page.

Project Timeline

Treatment of Stage I-III Squamous Cell Carcinoma (SCC) of the Anal Canal

Feb 7, 2023
Topic Initiated
Aug 18, 2023
Aug 20, 2024
Systematic Review
Page last reviewed September 2024
Page originally created August 2024

Internet Citation: Systematic Review: Treatment of Stage I-III Squamous Cell Anal Cancer. Content last reviewed September 2024. Effective Health Care Program, Agency for Healthcare Research and Quality, Rockville, MD.
https://effectivehealthcare.ahrq.gov/products/anal-cancer-treatment/research

Select to copy citation