Local tumor control and toxicity in HIV-associated anal carcinoma treated with radiotherapy in the era of antiretroviral therapy

Purpose To investigate the outcome of HIV-seropositive patients under highly active antiretroviral treatment (HAART) with anal cancer treated with radiotherapy (RT) alone or in combination with standard chemotherapy (CT). Patients and methods Clinical outcome of 81 HIV-seronegative patients (1988 – 2003) and 10 consecutive HIV-seropositive patients under HAART (1997 – 2003) that were treated with 3-D conformal RT of 59.4 Gy and standard 5-fluorouracil and mitomycin-C were retrospectively analysed. 10 TNM-stage and age matched HIV-seronegative patients (1992 – 2003) were compared with the 10 HIV-seropositive patients. Pattern of care, local disease control (LC), overall survival (OS), cancer-specific survival (CSS), and toxicity were assessed. Results RT with or without CT resulted in complete response in 100 % of HIV-seropositive patients. LC was impaired compared to matched HIV-seronegative patients after a median follow-up of 44 months (p = 0.03). OS at 5 years was 70 % in HIV-seropositive patients receiving HAART and 69 % in the matched controls. Colostomy-free survival was 70 % (HIV+) and 100 % (matched HIV-) and 78 % (all HIV-). No HIV-seropositive patient received an interstitial brachytherapy boost compared to 42 % of all HIV-seronegative patients and adherence to chemotherapy seemed to be difficult in HIV-seropositive patients. Acute hematological toxicity reaching 50 % was high in HIV-seropositive patients receiving MMC compared with 0 % in matched HIV-seronegative patients (p = 0.05) or 12 % in all HIV-seronegative patients. The rate of long-term side effects was low in HIV-seropositive patients. Conclusion Despite high response rates to organ preserving treatment with RT with or without CT, local tumor failure seems to be high in HIV-positive patients receiving HAART. HIV-seropositive patients are subject to treatment bias, being less likely treated with interstitial brachytherapy boost probably due to HIV-infection, and they are at risk to receive less chemotherapy.


Background
The incidence of cancer of the anal canal is rising due to the increasing prevalence of HIV-infection and HPV-infection [1][2][3][4]. Standard therapy for invasive anal cancer is radiotherapy (RT) or chemo-radiation resulting in local tumor control (LC) rates and disease-free survival (DFS) in HIV-seronegative patients approaching 72 % and 73 %, respectively [5][6][7]. Few data exist on treatment outcome in HIV-seropositive individuals. Retrospective survival analyses of cohort patients in the pre-HAART era indicate that HIV-infection is associated with poorer outcome after combined chemo-radiation [7][8][9][10]. Though, some investigators reported lower doses of RT and chemotherapy being applied in patients with HIV-infection [3,11]. Sideeffects tended to be more frequent and more intense in HIV-seropositive patients without HAART than in HIVseronegative patients in some reports [12][13][14] whereas in others acute toxicity was moderate [15]. The increased likelihood of therapy-related toxicity correlated with low CD4 count in HIV-seropositive patients in the pre-HAART era in one report [16].
The introduction of HAART resulted in an increase of CD4 counts in responders and prolongation of survival. The influence of HAART on concomitant cancer treatmentrelated toxicity and treatment outcome of anal cancer remains controversial. Analysing very small patient cohorts, some authors showed no changes of the overall survival (OS) rates of anal cancer since the introduction of HAART [1] while others reported favorable treatment and toxicity outcome compared with results of the non-HIV population [17,18].
The aim of this study was to investigate clinical characteristics of HIV-seropositive and HIV-seronegative patients and whether the outcome in respect of treatment toxicity and survival after standard curative 3-D conformal RT with or without chemotherapy (CT) of invasive cancer of the anal canal is comparable between HIV-seropositive patients receiving HAART and stage and age matched HIVseronegative patients.

Patients
Ninety-one patients presenting with histologically proven invasive carcinoma of the anal canal between 1988 and 2003 at the Department of Radiation Oncology, Zurich, were treated with curative 3-D conformal RT alone or combined with CT. First, clinical characteristics, pattern of care and outcome of 81 HIV-seronegative patients were retrospectively analysed. Then, 10 consecutive HIV-seropositive patients receiving HAART (1997 and were retrospectively compared to 10 HIV-seronegative patients (1992 -2003) matching for TNM-stage and age. Selection of matched HIV-negative patients was as follows: of 81 HIV-seronegative patients with invasive carcinoma of the anal canal, 42 patients matched for TNM-stage and of these 42 patients 10 patients corresponded for age and grading. After obtaining informed consent and internal institutional review approval, clinical outcome was analyzed by reviewing medical records and interviews of patients.
Pre-treatment staging was performed in all patients and included digital examination, endoluminal ultrasound, chest x-rays and either an abdominal ultrasound or CT scanning. Patients were staged according to the system adopted by the American Joint Committee on Cancer [19] and the Union International Contre le Cancer (UICC) before the primary treatment [19]. Post-treatment evaluation included a clinical examination including digital palpation at each visit and regular anal ultrasounds. Anoscopy and post-treatment biopsies were only performed when a suspicious lesion was identified.

Treatment
No patient in the HIV-seropositive had primary radical surgery with colostomy compared with 5 HIV-seronegative patients (6 %). All patients except one HIV-negative patient that died during treatment completed curative 3-D conformal RT. Standard RT was administered over a fiveweek period to a dose of 45 Gy in 1.8 Gy per fraction followed either directly by an external beam radiotherapy (EBRT) boost or an interstitial high-dose rate ( 192 Ir-HDR) boost after an interval of 2 -3 weeks to deliver a total dose of 59.4 Gy. The fractionated HDR brachytherapy boost (14 Gy/7 fractions, thrice daily) was applied to patients with T1 -T3 tumors appropriate for interstitial treatment after EBRT of 45 Gy.
Chemotherapy consisted of fluorouracil (5-FU) and mitomycin-C (MMC). 5-FU was applied as a continuous infusion during the first five days of radiotherapy at a dose of 750 mg/m 2 or over four days at 1000 mg/m 2 . The cycle was repeated during week five of RT. MMC was given as an i.v. bolus on day one of radiotherapy (15 mg/m 2 ) or twice during week 1 and 5 (10 mg/m 2 ). When contraindicated, MMC was replaced by cisplatin given i.v., during 1 hour infusion, in week 1 and 5. Criteria for "non-adherence to chemotherapy" included omittance of chemotherapy or dose-reduction because of side effects.

Toxicity
The common terminology criteria for adverse events v3.0 was used for assessing acute and late treatment toxicity [20]. Follow-up records addressing long-term toxicity were available for 95 % of the patients. Sphincter function was assessed by digital palpation.

Statistics
Mean values are indicated with standard deviation. Differences between groups on continuous variables were tested using the Mann-Whitney test. SPSS version 12 was used with exact p-values. Fisher's exact test was used to test for differences between groups on categorical variables. Survival was calculated from the beginning of RT to the day of death or the date of last follow-up and time-to-recurrence from beginning of RT to the day of recurrence or date of last follow-up. Survival curves for the two groups were plotted according to the Kaplan-Meier method. Differences in survival across the groups were tested using the Log rank (Mantel-Cox) test.

Results of HIV-seronegative patients
Mean age of the 81 HIV-seronegative patients was 61.6 +/ -12.7 years and 75 % of the patients were female (Table  1). Sixty-one patients (74 %) had a very good performance status (WHO 0°). Nine patients (11 %) had grad 1 and 3 patients (4 %) grade 2 WHO performance status and for 7 patients it was not known. At time of diagnosis, tumor stage distribution for T1/T2/T3/T4 was 15 %, 43 %, 31 % and 11 %. The majority of patients presented with nodal negative disease (61 %) and low grad tumors, i.e. G1 and G2 (62 %). Radical surgery with up-front colostomy was performed in 6 % of the patients. All of them received postoperative RT or chemo-radiation. RT-dose was 57.2 +/-5 Gy and RT-duration was 53.6 +/-17 days. Interstitial 192 Ir-HDR brachytherapy boost was appropriate in 42 %. CT was applied to 72 % and included MMC in 64 % of all patients. After a median follow-up of 45 months OS at 10 years was 48 % and CSS was 75 %. Local failure rate at 10 years was 13 %. Eleven patients received salvage surgery resulting in a 10-year colostomy-free rate of 84 % (78 % if the 5 patients with primary radical surgery are included). Two of them had not achieved complete response to chemo-radiation. Severe acute side effects were relatively rare with 31 %. The most frequent severe side effects were dermatitis (17 %), diarrhea (6 %) or thrombopenia (12 %). Two patients died during or immediately after treatment due to cerebral bleeding under thrombopenia or cardiac failure. Chronic side effects could be evaluated in 54 patients (67 %). Thirteen patients (24 %) experienced either ulcera (2 %), chronic proctitis (11 %) or sphincter pressure impairment (15 %).

Pattern of treatment
No HIV-seropositive patient received radical surgery with up-front colostomy. Also in the matched HIV-seronegative cohort no patient underwent primary surgery. Total RT dose did not differ between HIV-seropositive and matched HIV-seronegative patients (57 -58 Gy) whereas the duration of RT was longer in the matched HIV-seronegative cohort compared with the matched HIV-negative cohort (47.4 +/-6 d vs. 59.8 +/-10 d; p = 0.007). This was due to the interval between the EBRT and the brachytherapy boost used in HIV-seronegative patients. A significant difference was observed in the application of HDR-brachytherapy. No HIV-seropositive patient received HDRbrachytherapy boost compared with 60 % of the matched HIV-seronegative patients (p = 0.005). Adherence to chemotherapy seemed to be more difficult in HIV-seropositive patients with HAART than in the stage-and agematched HIV-seronegative patients (30 % vs. 80 %, respectively; p = 0.08). Lack of adherence was either due to non-compliance, co-morbidity or concomitant medica- tion, low CD4 counts, or severe thrombocytopenia before or during treatment.

Response and survival
Median follow-up was 44 months in the matched groups (HIV+: 8 -76 months; HIV-: 19 -116 months). Treatment with RT alone or combined with CT achieved complete remission in 100 % in HIV-seronegative as well as matched HIV-seropositive patients. Three HIV-seropositive patients with HAART suffered from local failure compared with no patient in the age-matched HIV-seronegative cohort, resulting in a statistically worse timeto-local recurrence (p = 0.03, Figure 3). Fifty-seven months after treatment, one HIV-seropositive patient suffered from a local relapse while no HIV-seronegative patient experienced local failure after 4 years. Colostomyfree survival was below 70 % (HIV+) and 100 % (matched HIV-). A non-significant trend towards impaired time-torecurrence for HIV-seropositive patients was observed (4 recurrences in HIV-positive patients versus 1 recurrence in HIV-negative patients; p = 0.1). In HIV-seropositive patients, local recurrences were histologically confirmed  in two cases. The two remaining patients suffered from locally extensive tumor disease (N = 1) or additional liver metastasis (N = 1). However, no significant difference in overall and cancer-specific survival was observed (Figure 1  Hematological toxicity: Three of six (50 %) HIV-seropositive patients receiving HAART who were treated with chemotherapy with MMC developed acute thrombocytopenia grade 3 (Table 1), compared to 0% of 8 matched HIV-negative patients. This resulted in a significant higher severe hematological toxicity rate (p = 0.05). Notably, two of the patients had CD4 count above 200 cells/μl.

Late toxicity
In respect of chronic side effects including chronic skin ulceration grade 3, complaints from chronic proctitis,  sphincter muscle dysfunction and incontinence grade 1 no significant difference was found between HIV-seropositive and matched HIV-seronegative patients. Prolonged wound healing was seen in only 1 patient with HIV-infection (10%) without evidence of tumor persistence compared with no patient in the matched HIV-seronegative cohort. Though, this patient developed a secondary lymphoma at the site of ulceration which was treated with chemotherapy. One patient of each group developed sphincter pressure impairment after treatment. The rate of sphincter function preservation was around 70 % (HIV+) and 100 % (matched HIV-). One HIV-seropositive patient under HAART died 2.5 years after RT because of sigmoid colon perforation, an area outside the RT treatment fields.
This study is the first matched-pair analysis on anal cancer patients with HIV-infection receiving HAART comparing outcome with HIV-seronegative controls accounting for the three most important prognostic-factors T-stage, Nstage and age [44][45][46]. Several findings are of clinical importance. (i) Standard 3-D conformal RT alone or combined with standard CT can achieve complete tumor response rates of 100% in patients with HIV-infection under HAART. (ii) Despite the good response rates and maintained treatment with HAART, local failure rate is significantly higher when compared with the prognostically similar tumor-stage and age matched control cohort possibly compromising colostomy-free and cancer-specific survival. (iii) Nevertheless, disease-free and overall survival is similar in HIV-positive patients under HAART compared with HIV-negative patients. (iv) Standard chemo-radiation with mitomycin-C is associated with a high rate of severe acute hematological toxicity but not with increased long-term side effects in HIV-seropositive patients with HAART. (v) HDR-brachytherapy was applied less frequently to HIV-seropositive patients and adherence to chemotherapy was more difficult.
Since the introduction of HAART, one comparative [18] and six non-comparative retrospective reports have evaluated the outcome of HIV-seropositive patients with access to HAART [1,17,[47][48][49][50]. Though, only 4 of these studies included more than 8 patients [1,17,18,50]. The study by Kinsella et al., presented (III and IV). Additionally, HIV-seronegative patients were an average of 17 years older than HIV-positive patients which may represent a different prognostic cohort. In respect of toxicity, they observed a low hematological severe toxicity rate in HIV-positive patients which is different to our study. This may be explained in part by the use of cisplatin instead of MMC. If so, cisplatin might be an advantageous alternative to MMC if applied at standard dose and not at highdose like in the study by Blazy et al. [50]. Another study by Allen-Mersh et al. reported treatment and toxicity outcome after chemo-radiation in 46 HIV-seropositive patients with sufficient response to HAART [17]. After a median follow-up of 35  There must be factors others than age or tumor stage for the higher local recurrence since the patients were well matched in respect of these prognostic factors. Such factors may be either patient-related or treatment-related. Patient-related factors may be persisting immunological alterations even after controlling HIV replication with HAART. Treatment-related factors may be the difficulties to apply full dose chemotherapy [52]. Chemotherapy and MMC in particular, was applied less frequently and with worse adherence in the HIV-seropositive patients either due to non-compliance, co-morbidity or concomitant medication, low CD4 counts, or severe thrombocytopenia before or during treatment. The importance of additional chemotherapy for the local tumor control has been demonstrated in several randomized trials [6,51,[53][54][55]. Thus, we assume that chemotherapy is important for the local treatment success in HIV-positive patients, as well.
Another difference of treatment between seropositive and seronegative patients was the use of brachytherapy that was more reluctantly applied in HIV-seropositive patients. The stringent matching procedure for TNM-stage, age and grade to some extent did not allow an additional matching for brachytherapy. Though, retrospective analysis of 81 HIV-seronegative patients at our Department revealed no better outcome for patients receiving a HDR brachytherapy boost compared to patients with an EBRT boost. Due to missing randomized trials or retrospective comparative studies indicative for a definitive clinical benefit of brachytherapy, we lack evidence to encourage the use of interstitial treatment in HIV-seropositive patients with anal cancer [56][57][58][59]. In the present cohort, it is not clear whether brachytherapy has been used more reluctantly because of the HIV-infection harbouring a risk of complications and treatment-related death or because of other reasons, such as the particular anatomical presentation of the disease.
There are some limitations to the present study. Because of the low incidence only a small number of patients could be analysed with limited statistical power. Therefore, only clear and important differences were found. Due to the limited number of cases, this study must be considered hypothesis-generating and not conclusive. The case-control analysis improves the validity of the study results although there may still be a selection bias and despite that the controls were derived through a stringent selection algorithm, which did not allow changing matching cases before analysis. Furthermore, when the HIVseropositive patients were compared with all, nonmatched consecutive 81 HIV-seronegative patients, timeto-local recurrence was still inferior in HIV-seropositive patients under HAART. Matching for T-/N-stage as well as age is important since both are significant statistical parameters and because the demographics of HIVpatients with anal cancer differ from those of HIV-seronegative patients a fact not considered in the study by Kinsella et al.. Grading is not an independent risk factor when adjusted for stage [60] and gender can not generally be considered as a prognostic factor, although some multivariate analyses suggest that women may have a better prognosis than men [51,[61][62][63]. The performance status was similar in all the patients investigated in this study, never below 80% of Karnofsky scale [46].
We conclude that HIV-seropositive patients with HAART and a good performance status are likely to achieve a complete response with standard 3-D conformal RT alone or combined with standard chemotherapy, although the risk for local relapse is high. Nevertheless, overall survival is comparable to HIV-seronegative patients. Despite good performance status the tolerability of chemo-radiation with mitomycin-C is limited by the toxicity of full dose chemotherapy. Therefore, in HIV-seropositive anal cancer patients with access to HAART, combined chemo-radiation is challenged to be more efficient at reduced toxicity at which cisplatin may represent a feasible option.

Competing interests
The author(s) declare that they have no competing interests.