2. Chemoradiation versus Radiation Therapy Alone
Organ preservation therapy for SCCA was borne out of the pioneering work of Norman Nigro, who first used preoperative chemotherapy and radiation therapy (RT) to convert unresectable patients to resectable and incidentally found high response rates at the time of surgery, prompting him to ultimately forego resection
[11][12]. Initial reports showed favorable outcomes with both RT alone as well as CRT. However, CRT was secured as the standard of care treatment for localized SCCA following two key trials, namely the United Kingdom Coordinating Committee on Cancer Research (UKCCCR) Anal Cancer Trial (ACT I) and the European Organization for Research and Treatment of Cancer (EORTC) trial
[13][14][15]. In ACT I, patients treated with CRT had a reduced risk of local failure and death from anal cancer compared to the RT alone arm
[13]. On long-term follow-up, CRT remained associated with a reduction in the risk of locoregional relapse, improved relapse-free and colostomy-free survival (CFS), and a reduction in the risk of dying from SCCA
[15]. The outcomes of the EORTC trial were similar, with CRT resulting in improved locoregional recurrence (LRR) rates and colostomy-free interval compared to RT alone
[14]. In ACT I, more acute toxicity was observed in the CRT group, with similar rates of late toxicity between the two arms
[13]. While nearly 40% of patients randomized on ACT I had T1–2 N0 disease
[16], patients with similar early-stage disease were excluded from the EORTC trial
[14] making decisive treatment decisions challenging for this patient population. It is important to highlight that patients included in ACT I underwent essentially clinical staging alone, potentially limiting any conclusions about treatment of early-stage patients included in this trial compared to contemporary treatment that incorporates modern imaging.
Given the limited number of patients with early-stage disease included in the randomized trials that defined CRT as the standard of care for localized SCCA, questions remain as to whether RT alone is sufficient treatment for the smaller proportion of patients with early-stage disease. A summary of select studies reporting on RT alone for early-stage SCCA is shown in . In a subset analysis of patients with T1–2 N0 disease included on ACT I, a clear advantage in treatment of these patients with CRT compared to RT alone was observed for local failure (RR = 0.49, 95% CI 0.29–0.71,
p = 0.0005)
[16]. However, multiple retrospective series have shown favorable results using RT alone, particularly in those with early-stage disease
[17][18][19]. A small series of 69 patients from 17 French institutions evaluated outcomes of patients with either Tis or T1 SCCA with tumor size ≤1 cm found on clinical exam or endosonography treated with RT or local excision (LE) alone
[20]. Of the 69 patients included in the study, three patients with Tis underwent LE alone, while the remaining 66 were treated with RT, with 26 of those patients undergoing LE before RT. Of the 66 patients who received RT, eight underwent brachytherapy alone. The 5-year OS, CFS, and DFS rates were 94%, 85%, and 89%, respectively.
Table 1. Select studies evaluating radiation therapy (RT) alone or comparing chemoradiation (CRT) to RT in early-stage anal cancer.
Several smaller retrospective studies have also reported on outcomes of CRT vs. RT in patients with early-stage SCCA. Fallai et al. reported on 62 elderly (age ≥70 years) patients treated with RT or CRT for clinically staged SCCA
[21]. Only 15% of patients included in the study were stage I, while 47% were stage II and 39% had stage III disease. Eight of the nine patients with stage I disease were treated with CRT with 5-year OS and locoregional control (LRC) rates of 100%. Overall, use of CRT was associated with improved 3-year outcomes, including DFS (85% vs. 46%,
p = 0.013), local control (LC) (80% vs. 60%,
p = 0.032), and LRC (81% vs. 61%,
p = 0.037), but not OS (85% vs. 67%,
p = 0.3). A review of 146 patients with T1–2 N0 SCCA treated at Geneva University Hospital in Switzerland was conducted to study the impact of concurrent chemotherapy on LRC and cancer-specific survival (CSS)
[22]. Staging was completed per the treatment period, with physical examination, abdominal ultrasound, and chest radiography used initially and incorporation of abdominopelvic computed tomography (CT), transrectal echoendoscopy, pelvic magnetic resonance imaging (MRI), and fluorodeoxyglucose positron emission tomography (PET) scans for patients treated more recently. Of the patients included in the study, 80% were T2 N0 and 48% were treated with RT alone, with 91% receiving split course treatment and over half of patients receiving a brachytherapy boost to the primary tumor. The mean ± SD tumor size in the RT and CRT groups was 2.9 ± 1.2 cm and 3.2 ± 0.9 cm, respectively,
p = 0.168. The 5-year rate of LRC for RT alone was 75.5% compared to 86.8% for patients treated with CRT,
p = 0.155. On multivariate analysis, treatment with CRT showed a trend toward significance for LRC (HR = 2.23, 95% CI 0.95–5.23,
p = 0.065). De Bari et al. reported on 122 patients with T1–2 N0 SCCA treated with CRT (70 patients) or RT alone (52 patients)
[23]. Similar to Zilli et al., imaging for staging was dependent on the treatment time period, with only 16% of patients undergoing staging pelvic MRI. Of note, only 29% of patients treated received prophylactic inguinal irradiation, and a brachytherapy boost to the primary tumor was delivered in 68% of patients. Delivery of CRT statistically improved LC (RR = 0.34, 95% CI 0.16–0.75,
p = 0.007).
The questionable benefit of CRT over RT alone for early-stage SCCA has also been investigated using large database studies. Buckstein et al. performed a cohort analysis comparing CRT to RT in elderly patients with stage I SCCA using the Surveillance, Epidemiology, and End Results (SEER) registry linked to Medicare
[3]. The final study population consisted of 99 patients treated with RT alone and 200 patients treated with CRT. Unadjusted analysis showed that patients treated with RT alone had inferior OS at 5 years compared to those receiving CRT (61% vs. 73%,
p = 0.002), but no difference in cause-specific survival, DFS, or rate of abdominoperineal resection was observed. Following propensity score matching, there was no significant difference in OS (
p = 0.08) and no significant difference in the other oncologic outcomes assessed. In the matched population, CRT was associated with a higher frequency of acute toxicity but no difference in late toxicity. We sought to determine if RT alone was sufficient for treatment of patients with stage I SCCA using the National Cancer Database (NCDB)
[4]. We identified 3552 stage I SCCA patients treated with CRT and 287 treated with RT alone. Patients treated with CRT were more likely to be ≥70 years old (33.1% vs. 19.7%,
p < 0.001) and less likely to be female (63.1% vs. 71.0%,
p < 0.001). Following propensity score matching, treatment with CRT compared to RT alone was associated with a 31% reduction in the risk of death (HR = 0.69, 95% CI 0.50–0.95,
p = 0.023).
Talwar et al. performed a systematic review and meta-analysis comparing RT to CRT for stage I SCCA patients
[5]. The authors included five retrospective studies with 3784 patients treated with CRT and 415 patients treated with RT. The 5-year OS was significantly higher for patients treated with CRT compared to RT alone (RR
= 1.18, 95% CI 1.10–1.26,
p < 0.00001). No difference in DFS between the groups was observed, although that endpoint was not reported in all of the studies included in the meta-analysis. Finally, of the studies able to report on toxicity, higher rates of both acute and late toxicity were reported in patients who received CRT.
3. Local Excision versus Chemoradiation
While the standard of care for localized SCCA is CRT, there has been some debate as to whether CRT is the optimal approach for patients with early-stage SCCA, particularly T1 N0 disease. This is because patients with T1 N0 disease represented a very small proportion of patients treated on the major randomized studies of SCCA
[13][14][24][25][26]. The number of series that report LC results of local excision alone are few, all with a small number of patients, some including T1 and T2 tumors, and all demonstrating local recurrence rates ranging 0–60%
[27][28][29][30][31]. In a recent retrospective study of 57 patients with T1 N0 SCCA, 13 received local excision alone and 44 received CRT
[32]. Local recurrences occurred in two of the 13 patients treated with LE (15% rate, both salvaged with surgery) and one of the 44 patients with CRT. There was no difference in 5-year progression-free survival (PFS) between the two cohorts (91% vs. 83%,
p = 0.57). A large retrospective study of patients with T1 N0 SCCA identified from the NCDB compared OS amongst 503 patients treated with local excision alone compared to 1740 patients treated with CRT from 2004 to 2012
[33]. The authors found that the use of local excision alone increased steadily during the study period from a rate of 17.3% in 2004 to 30.8% in 2012. The 5-year OS rate was similar in patients treated with local excision alone compared to CRT (85.3% vs. 86.8%,
p = 0.93). Interestingly, a comprehensive NCDB analysis of treatment outcomes in patients with stage I SCCA from 2004 to 2015 found slightly worse OS in patients treated with excision alone compared to CRT in a propensity-score-matched cohort (4-year OS 82.8% vs. 85.6%,
p = 0.045)
[4]. While these OS data from the NCDB studies are somewhat conflicting in a statistical manner, the OS rates of 83–85% for excision alone compared to 86–87% are numerically similar and likely not clinically significant differences. Therefore, excision alone can be considered for select patients with T1 N0 disease, which is supported by the NCCN guidelines
[34].
4. Modifications to Systemic Therapy
Radiation therapy with concurrent 5-FU and mitomycin C (MMC) remains the current standard of care for localized SCCA based on the results of multiple large randomized trials. The addition of MMC to 5-FU and RT was investigated in RTOG 87-04/ECOG 1289
[25]. Adding MMC improved the 4-year local failure rate (34% vs. 16%,
p = 0.0008), DFS (51% vs. 73%,
p = 0.0003), and CFS (59% vs. 71%,
p = 0.014) with no significant difference in OS (67% vs. 76%,
p = 0.31) but at the cost of increased grade 4–5 toxicity (8% vs. 26%,
p < 0.001). In an attempt to replace MMC, several studies compared replacing MMC for cisplatin with concurrent 5-FU and RT
[24][26][35]. RTOG 98-11 was a phase III randomized trial comparing 5-FU plus cisplatin induction chemotherapy followed by RT with concurrent 5-FU plus cisplatin to RT with concurrent 5-FU plus MMC
[24]. The initial results showed an improvement in colostomy rate with the use of MMC (10% with MMC vs. 19% with cisplatin,
p = 0.02), but no difference in 5-year LR, DFS, or OS despite higher rates of severe acute grade 3–4 hematologic toxicity with MMC (61% with MMC vs. 42% with cisplatin,
p < 0.001). At longer follow-up, DFS (67.8% vs. 57.8%,
p = 0.006) and OS (78.3% vs. 70.7%,
p = 0.026) at 5 years were statistically better for RT with concurrent 5-FU plus MMC compared to 5-FU plus cisplatin
[35]. In addition, a trend for improved CFS (71.9% vs. 65.0%,
p = 0.05), locoregional failure (20.0% vs. 26.4%,
p = 0.087), and colostomy failure (11.9% vs. 17.3%,
p = 0.074) with the addition of MMC was also reported. ACT II was a 2 × 2 factorial trial that investigated if replacing MMC with cisplatin improves response and if maintenance chemotherapy following CRT improves PFS
[26]. Patients were randomized to receive either MMC or cisplatin with concurrent 5-FU and RT with or without two additional courses of 5-FU and cisplatin. No difference in 3-year PFS was observed between the MMC and cisplatin groups (HR
= 0.95, 95% CI 0.75–1.19,
p = 0.63) or in those patients receiving maintenance chemotherapy. Of note, similar toxicity was observed in the MMC and cisplatin groups.
Modifications to this standard regimen have resulted in similar rates of efficacy with potential improvements in toxicity. The replacement of 5-FU by capecitabine has been investigated in multiple smaller phase II trials and retrospective reports. A multicenter phase II pilot study was performed in the United Kingdom that investigated the use of capecitabine delivered at a dose of 825 mg/m
2 twice daily during RT instead of 5-FU using the ACT II radiation regimen (50.4 Gy with concurrent MMC 12 mg/m
2 delivered on day 1)
[36]. The LC rate based on exam, imaging, or both at 6 months following completion of treatment was 90%. When evaluating compliance with treatment, 58% of patients completed both chemotherapy and RT as planned. Grade 3–4 acute toxicity was observed in 45% of patients. Oliveira et al. conducted a phase II trial consisting of 43 patient with either T2–4 N0 M0 or Tany N1–3 M0 SCCA treated with capecitabine 825 mg/m
2 twice daily during RT with a single dose of MMC 15 mg/m
2 on day 1
[37]. The primary endpoint of the study was LC at 6 months determined by clinical exam and imaging with either pelvic CT or MRI. The rate of LRC at 6 months was 86%. The main grade 3–4 toxicities were grade 3 radiation dermatitis (23.2%), grade 3 lymphopenia (11.6%), and grade 3 neutropenia (6.9%). Treatment interruption of capecitabine was required in 55.8% of patients for a mean duration of 11.2 ± 11.0 days due to primarily grade 3 radiation dermatitis and grade 2–3 hematologic toxicity.
Treatment in a larger cohort of patients receiving more modern treatment was conducted by Jones et al., who reviewed SCCA patients treated with intensity-modulated radiation therapy (IMRT) with concurrent MMC (single-dose on day 1) and either 5-FU or capecitabine from 50 centers in the United Kingdom
[38]. Data from 40 centers with 147 patients were included in the study, 35.4% of whom were treated with concurrent capecitabine/MMC and 64.6% treated with concurrent 5-FU/MMC. Although limited by available data, the 1-year relapse-free rates were not significantly different between the two groups (76.2% in the capecitabine/MMC vs. 79.3% in the 5-FU/MMC,
p = 0.80), nor were the 1-year CFS rates (77.5% in the capecitabine/MMC vs. 90.7% in the 5-FU/MMC,
p = 0.09). While no difference was observed in overall rates of grade 3 or 4 toxicity (45% for capecitabine/MMC vs. 55% for 5-FU/MMC), less grade 3–4 hematologic toxicity was observed in the capecitabine/MMC cohort (4% vs. 27%,
p < 0.001). A single institution series by Goodman et al. compared 107 patients with SCCA treated with IMRT and concurrent MMC (day 1 and 29 at 10 mg/m
2) with 5-FU (63 patients) or capecitabine (44 patients)
[39]. The 2-year oncologic outcomes between the groups were similar, including OS (87% for 5-FU vs. 98% for capecitabine,
p = 0.12), LRR (6.5% for 5-FU vs. 8.2% for capecitabine,
p = 0.78), distant metastasis (14.7% for 5-FU vs. 7.6% for capecitabine,
p = 0.26), and colostomy rate (5% for 5-FU vs. 9% for capecitabine,
p = 0.65). More grade 3 to 4 neutropenia was observed in the 5-FU group compared to the capecitabine group (52% vs. 20%,
p = 0.001). Treatment breaks due to toxicity were more likely in the 5-FU group compared to the capecitabine group (41% vs. 14%,
p = 0.006) with a median treatment duration significantly longer for patients receiving 5-FU (39 days, range 32–52 days vs. 37 days, range 32–44 days,
p < 0.001). Finally, dose reductions were more likely in the 5-FU group (52% vs. 16%,
p < 0.001). A meta-analysis of five trials reported pooled outcomes, including a complete response rate of 88% (83–94%) at 6 months post-treatment with an overall complete response rate evaluated at different time intervals of 91% (87–95%), with 93.5–100% of patients completing the planned RT dose
[40]. This result further highlights that capecitabine is an effective and convenient alternative to 5-FU for treatment of SCCA.
Dosing of MMC has varied between North American and European trials. While in North American trials, MMC has been dosed as 10 mg/m
2 delivered in two doses, European trials typically deliver 12–15 mg/m
2 on day 1 of RT
[13][14][24][25][26]. Reducing the dose of MMC conceivably would reduce toxicity. White et al. performed a single-institution retrospective review of 217 patients receiving definitive CRT for localized SCCA, comparing treatment outcomes and toxicity in patients who received one (154 patients) vs. two (63 patients) cycles of concurrent MMC
[41]. At 2-years, no significant differences in oncologic outcomes were observed between the two groups (one vs. two cycles), including PFS (78% vs. 85%,
p = 0.39), CSS (88% vs. 94%,
p = 0.11), CFS (87% vs. 92%,
p = 0.51), and OS (84% vs. 91%,
p = 0.16). Rates of grade ≥2 acute toxicity were higher in the two cycle group, including overall hematologic (89% vs. 73%,
p = 0.01), skin (97% vs. 84%,
p = 0.006), genitourinary (19% vs. 8%,
p = 0.04), and treatment-related death (5% vs. 0%,
p = 0.02). While limited, the results of this study indicate that a single dose of MMC is efficacious and may result in less acute treatment-related toxicity than two cycles. Of course, this result needs to be confirmed in prospective trials.