Primary surgery with or without postoperative radiotherapy in early stage squamous cell carcinoma in the anal canal and anal margin

Abstract

Background

Standard treatment of localized squamous cell carcinoma of the anus (SCCA) is radiotherapy (RT) combined with chemotherapy, that is, chemoradiation (CRT). Primary surgery has a limited role, but is a recommended treatment for small well differentiated SCCA localized in the anal margin, with re-excision or postoperative RT/CRT in case of involved surgical margins. The evidence supporting these strategies is limited.

Aim

To study the recurrence patterns and survival outcomes in patients treated with surgery alone compared with surgery followed by postoperative RT/CRT.

Material and methods

From a large Nordic database we identified 93 patients with stage TxT1–2N0M0 SCCA treated with surgery alone (n = 59) or surgery followed by RT/CRT (n = 34). Surgery consisted of local excision in 86 patients and abdominoperineal resection in seven patients, all of them in the surgery alone group. In 38 (41%) of the patients, the tumor was localized merely in the anal margin and in all remaining cases the anal canal was involved. Median RT dose to the tumor bed was 54 (range 46–66) Gy. Adjuvant RT to lymph nodes was given in 75% of the patients. Half of the patients received concomitant chemotherapy, usually 5-fluorouracil and mitomycin C.

Results

The locoregional recurrence (LRR) rate was significantly higher after surgery alone compared to surgery followed by adjuvant RT/CRT (36% vs. 9%, p = .006). The 3-year recurrence free survival (RFS) and overall survival (OS) were significantly better in patients who received postoperative RT/CRT than in patients who did not (3-year RFS 84.2% vs. 52.7%, p < .001 and 3-year OS 87.2% vs. 70%, p = .026).

Conclusions

Surgery alone of SCCA was associated with a high LRR rate and poor survival. The addition of postoperative RT/CRT lead to significantly improved locoregional control and survival.

Introduction

Squamous cell carcinoma of the anal region (SCCA), localized in the anal canal or the anal margin, is a rare type of cancer accounting for 1–2% of gastrointestinal malignancies [1]. The incidence is increasing world-wide. SCCA is strongly associated with human papilloma virus (HPV) infection [2]. The standard treatment for most patients with localized SCCA is radiotherapy (RT) with concurrent chemotherapy, so called chemoradiation (CRT). The most widely used regimen is a combination of 5-fluorouracil (5FU) and mitomycin C (MMC). CRT is an effective treatment, with cure rates of 60–80%, depending on tumor stage, but late side-effects are common [3–5].

Primary surgery has a limited role in the management of SCCA, but may according to ESMO and NCCN guidelines be a valid option in well differentiated tumors, smaller than 2 cm, localized in the anal margin. In case of involved surgical margins, re-excision or postoperative RT/CRT is recommended [1]. However, data supporting these strategies are very sparse. The outcome after surgery alone is only reported in a few small retrospective studies of older date [6–10].

From a large Nordic database we identified patients who underwent primary surgical resection for SCCA. The purpose was to describe recurrence patterns and survival in patients treated with surgery alone compared with surgery followed by postoperative RT/CRT.

Material and methods

Patients

This Nordic anal cancer database consists of 1266 patients diagnosed with SCCA between 1 July 2000 and 30 June 2007 in Norway, Denmark and Sweden. Of these, 886 patients were treated with RT or CRT according to Nordic guidelines, whereas 380 patients were not. Treatment results from the entire database have been previously reported [4]. We identified 93 patients with invasive SCCA, tumor stage TxT1–T2N0M0 who underwent primary surgery. Out of these, 59 patients were treated with surgery alone (group S), and 34 patients received postoperative RT or CRT (group S + RT/CRT) within 6 months after surgery (Figure 1). Data were retrieved retrospectively from the medical records and entered into the database. No assessment of the HPV status was done.

Figure 1. Patient selection flow diagram.

Tumor staging was performed according to institutional standards, with digital rectal examination and examination of inguinal lymph nodes, anorectoscopy, biopsy, and computer tomography (CT) of the abdomen and thorax as a minimum. The 4th edition of the UICC staging system served as basis for TNM classification, T1 indicates primary tumor ≤2 cm and T2 > 2–5 cm, respectively. The study was approved by ethics committees in all participating countries.

Treatment

Surgery consisted of local excision or abdominoperineal resection (APR). Local excision was performed with or without pre-surgical knowledge of SCCA, but detailed information regarding type of excision and preoperative knowledge of the diagnosis is lacking. The reasons for primary APR were previous radiotherapy in two cases (prostate cancer and uterus cancer), high age or comorbidities that disqualified them from standard RT/CRT. For resection radicality the following definition was used; R0 = microscopically radical with >1 mm margin, R1 = macroscopically but not microscopically radical, R2 = macroscopically not radical. The R-assessment was based on the original reports from the surgeon and pathologist, but no pathological reevaluation was performed.

In the S + RT/CRT group, RT was delivered to the primary tumor bed with a 1.5–3 cm margin. Elective lymph nodes were irradiated in most cases, depending on stage and localization, usually perirectal, presacral, iliac, and inguinal lymph nodes. Target volume definitions were based on CT scans. There were no pre-specified dose recommendations in the Nordic guidelines for the postoperative setting. The RT doses are presented below. RT was delivered with photon energy of 6–18 mV, five fractions per week, without gap. RT techniques varied from 2-field AP-PA to conformal methods with multiple fields.

Different chemotherapy schedules were used. The most common regimen was one cycle of 5FU and MMC during the first week of RT. A few patients received cisplatin and 5FU, either three cycles before RT or two cycles before and one cycle concomitant with RT. The doses of the different chemotherapeutic agents have been described in detail previously [4].

Follow-up assessments

After treatment, the patients were followed-up according to institutional routines, usually with clinical examination every 3 months the first 2 years and then every 6 months to 5 years. CT scans and biopsy were performed when clinically indicated. The median follow-up was 41 months.

Statistics

Data were summarized using standard descriptive statistics. Comparision between the treatments groups were done with Fisher’s exact test to compare categorical variables and with Mann–Whitney test to compare continuous variables.

Overall survival (OS) and recurrence free survival (RFS) rates were calculated using the Kaplan–Meier method and compared with the log-rank test. OS was calculated from the date of diagnosis to death from any cause. RFS was defined as the period from the date of diagnosis to loco-regional failure, distant failure, or death. Locoregional recurrence (LRR) was defined as a relapse in the anorectal area or pelvic or inguinal lymph nodes. Patients alive or lost during follow-up were censored. Cox regression model was used for univariate and multivariate analysis to evaluate the effect of different predictors influencing RFS. All statistical tests were two-sided and a p value ≤.05 was considered statistically significant. Data were analyzed using the SPSS version 22.0 (IBM, Armonk, NY) and R version 3.2.2 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Patient and treatment characteristics

Regarding baseline characteristics, patients in the S group were significantly older, had smaller tumors and a higher proportion of well to moderately differentiated tumors than patients in the S + RT/CRT group (Table 1). There were no significant differences between the two groups with respect to gender, WHO status, and localization.

Table 1. Patients and tumor characteristics.

	Surgery as the only treatment S	Surgery followed by RT/CRT S + RT/CRT	p Value
No. of patients, n	59	34	–
Age, years
Median	66	55.5	.026^a
Range	34–89	44–91	–
Gender, n (%)
Male	19 (32)	11 (32)	NS^b
Female	40 (68)	23 (68)	–
WHO, n (%)
0–1	46 (80)	28 (82)	NS^b
2–3	1 (2)	2 (6)	–
Unknown	12 (8)	4 (12)	–
Tumour size (mm)
Median	18	22	.026^a
Range	3–50	2–45	–
Missing, n	14	5	–
T stage, n (%)
TX	5 (9)	2 (6)	.008^b
T1	38 (64)	12 (35)	–
T2	16 (27)	20 (59)	–
Histologic differentiation, n (%)
Well	26 (44)	4 (12)	.002^b
Moderately	15 (25)	13 (38)	–
Poorly	5 (9)	8 (24)	–
Unknown	13 (22)	9 (26)	–
Localization, n (%)
Margin	23 (39)	15 (44)	NS^b
Canal	30 (51)	16 (47)	–
Margin and canal	5 (9)	3 (9)	–
Unknown	1 (1)	0 (0)	–

^a Mann–Whitney test.

^b Fisher’s exact test.

NS: non-significant.

In the S group, 52 (88%) patients were operated by local excision and five (12%) underwent an APR, whereas all patients in the S + RT/CRT group had a local excision (Table 2). Regarding surgical radicality, there was a significantly higher percentage of radical resections (R0) in the S group compared to the S + RT/CRT group, 68% vs. 12% (p < .001).

Table 2. Primary treatment.

	Surgery as the only treatment S	Surgery followed by RT/CRT S + RT/CRT
No. of patients, n	59	34
Surgery performed, n (%)
Local excison	52 (88)	34 (100)
APR	7 (12)	0 (0)
Radicality, n (%)
R0	37 (68)	4 (12)
R1	17 (29)	25 (74)
R2	0 (0)	3 (9)
Unknown	5 (3)	2 (5)
Radiotherapy, n (%)
Primary tumor	0	34 (100)
Adjuvant ln	–	25 (75)
Brachytherapy	–	1 (3)
Chemotherapy, n (%)		17 (50)
Cisplatin/5FU	0	3 (18)
5FU/MMC	–	14 (82)

For patients in the S + RT/CRT group, the median external RT dose was 54 Gy (range 46–66 Gy) to the tumor bed. Twenty-five patients (75%) received adjuvant lymph node irradiation (100% iliacal and 84% inguinal) to a median dose of 46 Gy (range 26–46 Gy). Sixteen patients (47%) were treated by RT alone, one patient (3%) had external RT combined with brachytherapy and 17 patients (50%) received CRT. The most commonly used regimen was a combination of 5FU and MMC (82%). Three patients (18%) received 5FU with cisplatin (Table 2).

Patterns of recurrence

There were significantly more LRRs in the S group than in the S + RT/CRT group, 36% vs. 9% (p = .006). Seventeen patients (29%) in the S group recurred in the anal region, whereas no anal relapses were seen in the S + RT/CRT group. The only recurrences in the S + RT/CRT group occurred in regional lymph nodes. Two out of those three patients did not receive prophylactic inguinal irradiation (Table 3).

Table 3. Locoregional recurrence.

Recurrence site	Surgery as the only treatment S	Surgery followed by RT/CRT S + RT/CRT	p Value^a
No. of patients, n	59	34	.006
Locoregional, n (%)^b	21 (36)	3 (9)	–
Anus	17 (29)	0 (0)	<.001
Lymph-nodes	8 (14)	3 (9)	.74
Inguinal	5 (9)	3 (9)	–
Iliacal	3 (5)	0 (0)	–
Distant, n (%)	4 (7)	0
Lymph-nodes	1	–	–
Lung	2	–
Liver	1	–	–
Bone	0	–	–
Other	1	–	–
Treatment of locoregional recurrence, n	21^c	2^c
Surgery	11	1	–
Radiotherapy	13	2	–
Chemotherapy	6	0	–

^a Fisher’s exact test.

^b Some patients had recurrence both in the anus and lymph nodes.

^cSome patients had more than one treatment modality for recurrence.

In the S group, the LRR pattern was analyzed by subgroups (Table 4). Out of the 23 patients with a tumor involving anal margin only, seven patients (30%) had a LRR, six of them after R0 resection. Forty-three percent of patients whose tumor involved the anal canal had LRR, nine of them after R0 resection. A total of 37 patients had an R0 resection and 14 of them (38%) had LRR.

Table 4. Locoregional recurrence in the surgery alone group.

	Number of patients n = 59	Number of recurrences n = 22	%
T stage
Tx	5	3	60
T1	38	12	32
T2	16	7	44
Localization
Margin only	23	7	30
Canal	35	15	43
Unknown	1	0	0
Type of surgery
Local excision	52	19	37
APR	7	3	43
Radicality
R0	37	14	38
R1–2	17	5	30
Unknown	5	3	60
Histological differentiation
Well-moderately	41	14	34
Poorly-undifferentiated	5	2	40
Unknown	13	6	46

Special analysis was performed in a subgroup of patients in the S cohort, basically fulfilling the criteria ESMO guidelines recommending surgery alone. We thus identified 11 patients with well differentiated tumors ≤2 cm in size, localized in the anal margin with R0 resection and four of them (36%) had anal recurrence.

Out of the seven patients in the S group that were operated with APR, four (57%) developed recurrence, three of them LRR, and one patient had liver metastases.

All 21 patients in the S group who developed LRR were subjected to salvage treatment (Table 3), seven patients with surgery alone, nine patients had RT/CRT alone, and four patients had both RT/CRT and surgical resection.

Survival outcomes

There was a significantly worse 3-year RFS (Figure 2(a)) in the S-group than in the S + RT/CRT group, 52.7% vs. 84.2% (log-rank p < .001). The 3-year OS (Figure 2(b)) was also significantly inferior in the S-group than in the S + RT/CRT group, 70.0% vs. 87.2% (log-rank p = .026).

Figure 2. Recurrence-free (a) and overall (b) survival after surgery alone (S) and surgery followed by postoperative radiotherapy/chemoradiation (S + RT/CRT), respectively.

Uni- and multivariable analysis with Cox regression were conducted to evaluate the impact of age, gender, T stage, localization, histological differentiation, radicality, and RT/CRT on RFS. In the multivariable analysis, the addition of RT/CRT was the only factor with significant influence on RFS (Table 5).

Table 5. Predictors for recurrence-free survival after primary surgery.

		Univariate analysis		Multivariable analysis
Parameter	n	HR	p^a	HR	p^a
Age
≤60	42	1	.018	1	.09
>60	51	2.54	–	2.0	–
Sex
Male	30	1	.51	1	.51
Female	63	0.79	–	0.80	–
T stage
T1/Tx	57	1	.19	1	.87
T2	36	0.62	–	0.93	–
Localization
Margin only	38	0.84	.63	0.83	.62
Canal	55	1	–	1	–
Histologic differentiation
Well-moderately	35	1	.11	1	.53
Poorly-unknown	58	0.54	–	0.78	–
Radicality
R0	41	1	.059	1	.50
R1/R2/unknown	52	0.48	–	0.78	–
RT/CRT
Yes	34	0.21	.00028	0.29	.023
No	59	1	–	–	–

RT: radiotherapy; CRT: chemoradiation.

^a Cox regression.

Discussion

This study demonstrated a high local failure rate and a poor survival in patients treated with surgery alone for early squamous cell carcinoma of the anal canal or anal margin. Patients that received postoperative RT or CRT had a significantly better local tumor control and survival compared with those treated with surgery alone.

SCCA is a rare malignancy that historically was treated by surgery only, until Nigro et al. reported high response rates with low intensity CRT [11]. Based on subsequent studies, CRT has been established as the therapeutic mainstay for most patients with SCCA. Surgery is used as salvage treatment for patients with residual or recurrent tumors [3,12,13].

Primary surgery has a limited role in the management of SCCA, but according to ESMO guidelines it is a valid treatment option for patients with small (<2 cm) well differentiated tumors of the anal margin, followed by re-excision or RT/CRT if the resection margin is ≤5 mm [1]. However, the evidence supporting these strategies is rather weak. Regarding the efficacy of surgery alone there are a few reports in the literature, mostly based on patients treated in the 1950–1970s, all of them retrospective reviews with heterogenous populations. Schraut et al. [6] showed good results after local excision of predominantly carcinoma in situ or microinvasive perianal tumors, 2 cm or less in diameter, but a high failure rate for deeply invasive lesions. Similar results were presented by Boman et al. [7], with successful outcome in 12 out of 13 patients treated with local excision for superficially invasive tumors in the anal canal, but considerably worse outcome after local excision of more deeply invasive tumors. Longo et al. [8] demonstrated local recurrence in four out of nine patients (44%) with invasive tumors in the anal canal stage T1N0 and in three out of three patients (100%) with stage T2–3N0, after local excision alone. Greenall et al. [9] found 88% cause-specific 5-year survival after local excision of tumors in the anal margin, but a locoregional recurrence rate of 46%. Treatment outcome divided by depth of invasion was not presented in that study. High locoregional failure rates were also reported by Goldman et al. [10], in four out of seven (57%) patients treated with APR alone and in seven of nine (78%) patients treated with local excision alone for a stage T1–2N0 cancer in the anal canal. In a more recent study [14] based on the National Cancer Database, 2243 patients with stage T1N0 SCCA localized in the anal canal were analyzed. Five hundred and thirteen patients were treated with local excision alone and the remaining 1740 patients received CRT. A high 5-year OS rate of 85–87% was found, with no difference between the treatment groups. However, no data were presented regarding LRR or the outcome for patients that had undergone local excision prior to their CRT. They also showed that the proportion of patients treated with local excision alone has increased with time in the US, especially for tumors ≤1 cm. This suggests that these very small tumors may be managed adequately with local excision alone, but the majority of patients have a much larger tumor at diagnosis. To detect a substantial portion of SCCA while they are still less than 1 cm in size, requires screening programs, which have been implemented for high-risk populations at some centers in, for example, the US.

Thus, it seems like good results with surgery alone were achieved mainly in patients with superficially invasive tumors [6,7], that is, what today would be classified as carcinoma in situ or high-grade anal intraepithelial neoplasia, for which local excision is still the treatment of choice. Regarding the more invasive SCCAs, high local recurrence rates have been reported [7,8,10], well in accordance with our results, showing a 36% locoregional failure rate, but also a poor 3-year OS of only 70%, after surgery alone.

Among the patients in our study who received postoperative RT/CRT, using a median RT dose of 54 Gy, ranging from 46 to 66 Gy, the oncological outcome was clearly better than after surgery alone. LRR occurred in three out of 34 patients (9%), none of them with recurrence in the anal region. There was also a significantly improved RFS and OS in the S + RT/CRT group compared to the S group. These results are in line with previous studies in the literature. Ortholan et al. [15] reported local failure in three out of 23 patients treated with local excision of a ≤10 mm tumor in the anal canal, followed by RT and the authors recommend an RT dose of 40–50 Gy after R0/R1 resection and 50–60 Gy after R2 resection. Hatfield et al. [16] demonstrated LRR in only one out of 21 patients who received RT 30 Gy to the tumor bed, without elective lymph node irradiation, and one cycle of 5FU/MMC after local excision of a T1-T2N0 tumor in the anal canal or anal margin. Hu et al. [17] showed an 83% local control rate among 14 patients who went through R0/R1 excisional biopsy followed by CRT, with no apparent difference in local control whether RT doses of 30–34 Gy or 45–50.4 Gy were given. Thus, postoperative RT/CRT seems to be effective. Regarding optimal RT dosing in this setting no firm conclusions can be drawn. In the present study, RT doses ranging from 46 to 66 Gy were used, that is, higher doses than in most previous series. Based on this literature surveyit appears that 30–40 Gy to the tumor bed could be sufficient in the majority of cases.

Whether elective lymph node irradiation should be given or not is unclear. Hatfield et al. [16] showed low LRR rates despite omission of elective lymph nodes. From the studies by Ortholan et al. [15] and Hu et al. [17] the elective lymph node targets applied in the postoperative setting were not described in detail. In our study, adjuvant lymph nodes were irradiated in 75%, resulting in a very low frequency of lymph node recurrences. In a previous study from our group [4], we found an 11% risk of lymph node recurrence in patients with a T1N0 anal cancer treated with definitive RT without elective lymph node irradiation. At our institution elective lymph nodes are routinely included in the radiation volumes, both in the definitive and postoperative setting.

A limitation of the current study is the retrospective design and its limited sample size, but with 93 patients it is still one of the largest studies to-date on patients treated with primary resection for an early SCCA in the anal region. No previous study has compared the results after surgery alone with surgery followed by RT/CRT ‘head-to-head’. Even though this was not a randomized comparison we feel that our study may contribute with valuable information. The baseline characteristics differed somewhat between the treatment groups. Patients in the S + RT/CRT group were significantly younger, but had larger tumors with less favorable histological differentiation (Table 1) and were more frequently R1–2 resected (Table 2) than patients in the S group. Despite these tumor-related disadvantages, the outcome was clearly better in the S + RT/CRT group, indicating that postoperative RT/CRT is an effective strategy.

According to ESMO guidelines, surgery alone is contraindicated for tumors localized in the anal canal [1]. This notion was confirmed in the present study showing a 43% LRR rate for this subgroup (Table 4). On the contrary, surgery alone is the recommended treatment for patients with small (<2 cm) well differentiated tumors of the anal margin that have been radically excised with a resection margin of >5 mm. In the present study, 11 patients were radically excised for a ≤2 cm tumor in the anal margin and four of them (36%) developed LRR. This raises the question whether postoperative RT/CRT should be offered to all patients after primary excision, based on the general conclusion from our study and previous reports.

The present results indicate that the addition of postoperative RT/CRT improves the locoregional control and the overall survival after primary surgery. However, RT/CRT against the anal region may cause severe late side-effects, such as fecal incontinence, sexual dysfunction, and musculo-skeletal disorders [5,18]. Therefore, the possible benefits of the treatment must be weighed against the risk of treatment sequele, but if a wait-and-watch strategy is chosen after the resection, both the physician and the patients must be aware of the high risk of recurrence and a tight follow-up schedule is recommended.To further elucidate these issues a prospective non-randomized phase II study (PLATO-ACT 3) has been initiated in the UK, in which patients who have undergone an R0 resection (with >1 mm margin) will receive no further treatment whereas patients with an R1/R2 resection will be treated with postoperative 41.4 Gy in/23 fractions. RT to the tumor bed, combined with MMC and Capecitabine. It should be noted that the resection margin cutoff used in the PLATO-ACT 3 study is different from the 5 mm that is recommended in the current ESMO guidelines [1].

Since the optimal management is unclear, all patients who have undergone primary surgery for a SCCA should be discussed at a multidisciplinary team including surgeons, oncologists, pathologist, and radiologist.

Conclusions

Surgery alone of SCCA was associated with a high LRR rate and poor survival. The addition of postoperative RT/CRT lead to significantly improved locoregional control and survival.

Aknowledgments

Many thanks to NOAC (Nordic Analcancer) group members for the collection of clinical data.

Disclosure statement

None of the authors have any conflicts of interest to declare.

Additional information

Funding

The study was supported by Skåne Regional Council, Sweden.