Complementary and Salvage Surgeries in the Treatment of Cervical Cancer Persistent Following Chemo Radiation. Experience with 105 Patients.

Objective: The purpose of this review is to show results of surgical treatment performed 6-12 weeks after chemo radiation, with or without brachytherapy, in patients with locally advanced cervical cancer with documented or suspected tumour. Material and Methods: This was a retrospective study of patients with cervical cancer treated in the Oncology Service of Hospital General of Mexico during the period 2008-2018 who underwent surgery to perform complementary or salvage surgery due to actual or perceived tumour persistence. Results: Patients had a median age of 46.5 years (range= 25-71) and underwent 105 surgeries which included 75 hysterectomies with or without lymphadenectomy (71.4%) and 30 pelvic exenterations (28.5%). Major complications occurred in 7/75 (9.3%) of hysterectomies and 9/30 (30.0%) exenterations. In this series, there was no postoperative mortality. In 105 hysterectomies and pelvic exenterations, the Kaplan Meier curves were generated for disease-free survival (DFS) and overall survival (OS) with a median follow-up of 20 months and 24.0 months respectively. Common variables for the 105 hysterectomies and exenterations demonstrated by univariate analysis for (DFS) and (OS) significant differences for age < 35 years (p=0.031 and p=0.051), adenocarcinoma (p=0.023 and p=0.019) and for the absence of lymph vascular infiltration on surgical specimens (p=0.003 and p=0.001). The multivariate analysis for DFS and OS only showed significant differences for the absence vs presence of lymph vascular invasion (p=0.001 and p=0.002). Seventy -five patients treated with hysterectomies had DFS with a mean of 29.4 months and OS of 32.4 months. In the multivariate analysis, Log Rank test showed significant differences for the DFS and OS in favour of class 3 hysterectomy (p<0.0001 and p=0.003) and DFS for the absence of lymph node metastases (p = 0.059). In patients with exenterations, the multivariate analysis for OS demonstrated a significant difference with parametrial infiltration (p=0.012). In this series 13 of 105 (12.5%) patients developed recurrences with 53.8% % being distant. Conclusions: Post chemo radiation surgery in patients with suspected or documented persistent cervical cancer showed better prognosis for adenocarcinoma as compared to squamous carcinoma and less morbidity in the absence of brachytherapy. The high number of distant metastases in patients with tumour recurrence suggests the importance of administering adjuvant chemotherapy to these at-risk patients.


Introduction
Cervical cancer is a serious health problem in developing countries due to its presentation in advanced stages and is the leading cause of death from gynaecological neoplasms [1][2][3]. The World Health Organization (WHO) estimated in 2020 there will be 105,963 new cases of cancer in women in Mexico, of which 20,691 (19.5%) appeared in the genital tract. Cervical cancer leads the list of gynaecologic cancers in terms of incidence and mortality, with 9,439 new cases (45.6%) and 4,395 deaths (49.6%) [4]. In Mexican institutions such as the National Cancer Institute and the General Hospital of Mexico (GHM) [5,6], the number of affected patients with cervical cancer is only surpassed by those with breast cancer. Cervical cancer is predominantly diagnosed in advanced stages of the disease for which the treatment of choice is a combination of pelvic radiation with concomitant platinum agentbased chemotherapy followed by the application of Intracavitary Brachytherapy (ICBT) [7,8]. Although Chemo radiation (CCRT) maintains its validity for the management of locally advanced cervical cancer, therapeutic schemes have been published that include adjuvant surgery for some of these patients [12][13][14]. This increases the possibility of adapting the established treatments in order to improve specific oncologic results with conventional therapy. In 2003, the Gynaecologic Oncology Group of the United States of America [12] published data that patients with tumours of 4-6 cm limited to the cervix benefited from better local control of the disease when hysterectomy was added to the pelvic radiation (RT). In 2010, Houvenaeghel et al [13] published data on 35 patients with advanced lesions operated with hysterectomies or pelvic exonerations after receiving concurrent chemo radiotherapy (CCRT). Loco regional control was obtained in 88.6% of patients with a 10-year disease-free survival (DFS) of 66.4%. The authors concluded that adjuvant surgery reduces the risk of local recurrences and increases DFS. Fanfani et al [14]  results obtained with CCRT vs. CCRT plus hysterectomies in patients with stage III CC and reported a significant increase in local recurrences in the first group (p = 0.0210), a lower number of deaths in patients undergoing surgery (p = 0.021) and a 3-year DFS of 62.9% and 68.3% respectively (p = 0.0686). Failure to administer intracavitary brachytherapy represents incomplete treatment for patients undergoing CCRT. This deficiency constitutes a reason to consider increasing the external beam radiotherapy (EBRT) dose or to evaluate complementary surgical treatment. Walji et al [8] evaluated both treatments in 19/208 patients in whom incomplete ICBT was administered. Fourteen received an EBRT overdose (7 recurrent and 6 died of tumour activity) and five underwent hysterectomies class 1 or 3, all of whom experienced DFS at 5 years [15]. Recently the radiotherapy unit of our institution underwent a remodelling process during which it was not possible to administer ICBT in patients with cervical cancer who required concurrent chemo radiotherapy (CCRT). Some were channelled to other hospitals to complete their treatment and other patients were considered for complementary surgeries to conclude their treatment. In this setting, we present our experience in 62 patients who underwent surgery which included 56 hysterectomies with or without pelvic lymphadenectomy, 3 exenterations and 3 laparotomies for unresectable tumours. In 3/62 (4.8%) patients, there was a major complication with no operative mortality. An average follow-up of 30 months without tumour recurrence was obtained in 76.2% of all cases. In a subgroup analysis, a 30-month DFS period was obtained in 34/40 (85%) patients vs. 8/13 (61.5%) patients in whom there was a report of tumour activity (p = 0.03). DFS was noted in 12/13 (92.3%) patients in clinical stage I, 27/33 (81.8%) in stage II and 7/14 (50%) in stage III. It was concluded that surgery can be used as an alternative adjuvant treatment for patients with cervical cancer who require CCRT for its management and who cannot receive ICBT [16]. The aim of this publication is to present the results of surgical treatment in 127 patients with locally advanced cervical cancer whose tumour persistence was documented by biopsy, incomplete chemo radiation treatment due to lack of brachytherapy, clinical suspicion and/or by imaging studies when these were available after CCRT or with histopathology of poor prognosis (advanced stage adenocarcinomas) [17][18][19].

Material and Methods
This is a retrospective study of patients treated in the Oncology Unit of GHM during the period from January 2008 to December 2018 with the diagnosis of locally advanced cervical cancer (either squamous cell carcinoma (SCC) or adenocarcinoma) who received treatment with conventional or incomplete CCRT: External Beam Radiation Therapy 50-54 Gy with linear accelerator: Intensity-modulated radiation therapy (IMRT) or volumetric arc therapy (VMAT) + chemotherapy with Cisplatin or carboplatin at conventional doses. Intracavitary radiotherapy (brachytherapy30 Gy) high 3D rate with Iridium 192 or low dose rate with Cesium 137, or Tele therapy 50-54 Gy plus Chemotherapy with Cisplatin or Carboplatin weekly. The patients were laparotomies 6-12 weeks after concomitant chemo radiation treatment was completed under the following criteria: 1. chemo radiation without brachytherapy, 2. the presence of residual tumour in the cervix and or vagina documented by biopsy, 3. residual tumour in parametria by imaging studies (CT and PET), when available, 4. a clinical suspicion of residual tumour in parametria made by the specialty's professors in the absence of imaging studies; 5. reports of adenocarcinomas in advanced stages in which a less favourable response to conventional treatment has been documented [17][18][19]. Surgeries were performed according to the following preoperative and trans operative findings [9][10][11][12][13][14][15]. Extra facial hysterectomy class 1, [9,12] on suspicion of residual tumour in the cervix and / or vagina for not having received brachytherapy; Class 2 hysterectomy, (Stage IB1:FIGO-2018, in treatment-virgin patients) [3]: central persistence < 2 cm with resection of the internal third of the parametrium without pelvic lymphadenectomy and no suspicion of lymph node metastases or lymphadenectomy with suspected lymph node activity; Class 3 hysterectomy, [9,10,14] advanced stage neoplasia with infiltrating central residual tumour of 3 or >4 cm. Pelvic exenteration for neoplasms with residual tumour in the parametrium or with inability to separate the vesicovaginal or rectovaginal septa in central neoplasms. Pelvic exenterations (PE) were divided as follows: anterior (with Bricker duct), posterior (with resection of the rectum plus definitive colostomy) or total (with Bricker duct or sigmoid duct plus definitive colostomies). The major surgical morbidity was considered to be the complications obtained during the first 30 days following surgery. This included anastomosis dehiscence, bladder dysfunction, pneumonia and intestinal obstruction. The variables analysed included age, type of surgery performed, history of having received chemo radiation (CCRT), clinical stage and pathology reports with or without local or regional residual disease in the surgical specimen (cervix, vagina, uterine body, parametrium), metastases to the adnexa and report of lymph node metastases. For pelvic exenterations, the following variables were also analysed: report of central residual tumour defined as tumour limited to the remains of the cervix and/or vagina, diagnosis of parametrical invasion with or without infiltration to the pelvic wall, histologic evidence of invasion of the bladder or rectum and report of hydronephrosis previously observed through preoperative imaging studies. For age, we selected the analysis of the groups 35 years old and under and 36 years old and greater because a previous publication on prognostic factors for pelvic exenterations for cervical cancer demonstrated that patients aged 35 years and younger had a better prognosis [20]. The clinical classification used was that of the International Federation of Gynaecology and Obstetrics (FIGO) in its 2018 version [3] with consideration for stage I, IB2 and IB3 lesions, Stage II, IIA and IIB lesions and for stage III, IIIA and IIIB neoplasms. For the analysis of the variables, descriptive statistics were used with calculation of means and proportions for numerical and categorical variables. Survival analysis was performed using the Kaplan-Meier method. Differences between groups were assessed using the log-range test. Univariate and multivariate Cox proportional hazards models were used to identify variables that correlated with DFS and overall survival (OS) with 95% confidence intervals being calculated for the odds ratio. All statistical analyses were performed with the SPSS 22.0 statistical program.

Residual Tumour
Forty-six of 105 (43.8%) patients who were operated upon for curative purposes had a complete response to radiotherapy without residual tumour in the surgical samples.

Surgical morbidity and mortality
An evaluation of postoperative complications was recorded according to the treatment prior to the surgery and the surgical procedures performed. For the patients who received concurrent chemo radiotherapy, the figures were 15/83 (18.0%) and 1/22 (4.5%) for those with chemo radiation without brachytherapy (p = 0.2912). Major complications were present in 7/75 (9.3%) hysterectomies and 9/30 (30.0%) exenterations. In this series, there was no postoperative mortality.

Complementary or salvage hysterectomies
Seven of 75 patients (9.3%) who had received CCRT had major postoperative complications. Of these, 4/12 (33.3%) were present in class 3 hysterectomies (2 rectovaginal fistulas that required a definitive colostomy, 1 ureterovaginal fistula conservatively managed and 1 with bladder dysfunction which resolved with conservative therapy). In addition, 1 of 46 (2.3%) class II hysterectomies experienced an intestinal obstruction requiring surgery and 2/17 (11.7%) patients with class 1 hysterectomies developed rectovaginal and vesicovaginal fistula requiring a colostomy and ileal duct respectively.

Results of surgical treatment
The Ninety-two patients (87.6%) had follow-up from 1 to 79 months without evidence of tumour activity at their last evaluation. At 60 months, the figures for DFS and OS were 8.6% and 9.6%, respectively. For DFS and OS, the univariate analysis showed significant differences in favour of age < 35 years (p=0.031 and p=0.051), adenocarcinoma (p=0.023 and p=0.019), and the absence of lymph vascular infiltration (p=0. 003   Survival analysis was performed using the Kaplan-Meier method.
Differences between groups were assessed by using the log-rank test. Only variables that were statistically significant in the univariate analysis were analysed in multivariate Cox regression models. Only variables that were statistically significant in the univariate analysis were analysed in multivariate Cox regression models.

Complementary or salvage hysterectomies
General characteristics of 75 patients undergoing hysterectomies are shown in Table 4.   Figure 2.
The mean OS was 32.4 months and 38 (56.7%) patients had an OS of >25 months without evidence of disease. Figure 2.       Twenty-five of the 30 patients (83.3%) had a follow-up of 1 to 79 months without evidence of disease with a mean DFS of 18.7 months and OS of 20.2 months Figure 5.
In this series 7/25 (28.0%) patients had a disease-free follow-up of >25 months. The remainder ceased follow-up without evidence of disease during the course of the first 2 years. The univariate analysis for DFS showed variables with statistical significance for prognosis: absence of residual tumour in the surgical specimen (p=0.034) and the report of central tumour (p=0.024) and parametria infiltration (p=0.007).   In the multivariate analysis for DFS, no statistically significant differences were found.   Differences between groups were assessed using the Log-Rank Test

Tumour recurrences
The number of recurrences was 13/105 (12.3%) during the period of 5 to 29 months after surgery with an average of 12.3 months. This corresponded to 9/75 (12.0%) hysterectomies and 4/30 (13.3%) exenterations. Ten out of 74 (13.5%) patients had a previous history of radiotherapy plus concomitant chemotherapy and 3/31(9.6%) had tele therapy plus chemotherapy. Five out of 105 patients (4.7%) with radical surgery with no residual tumour in the surgical specimens developed tumour recurrences. The includes 3/75 (4.0%) of the hysterectomies and 2/30 (6.6%) exenterations. In 6/13 cases (46.1%), the recurrences were local or loco regional and distant in 7 (53.8%). In 3 cases (23.0%), the dissemination was regional and distant and the remaining 4 (30.7%) had several distant sites without loco regional activity. We performed additional therapeutic procedures in two patients, both stage IIB adenocarcinoma who received CCRT and class 3 hysterectomy. Case 1: the patient had been treated with incomplete CCRT, had residual tumour in the sample with the neoplasm recurring in the vaginal dome and required treatment with adjuvant ICBT at 30 Gy with a DFS of 36 months. Case 2: the patient had a local residual tumour and a metastatic lymph node and 7 months later had tumour recurrence in the vaginal dome and spleen requiring treatment with 4 cycles of paclitaxel plus vinorelbine and 5 cycles of gemcitabine with complete response demonstrating 24-month follow-up without tumour activity.

Discussion
The role that complementary surgery in patients with local advanced cervical cancer previously treated with CCRT is considered controversial due to surgical morbidity, mortality and oncologic results at 5 years compared to conventional treatment scheme with CCRT alone [11,12,14,21]. Working in an institution without optimal resources for the diagnosis and conventional treatment of advanced CC and the fact that the patients are admitted with advanced neoplasms, lack social security and live far from Mexico City, makes it very difficult to keep them under surveillance. Therefore, our Service decided to implement the treatment plan shown here for patients who received incomplete CCRT [8] or in patients who did receive it but who ended up with a confirmed or suspected residual tumour and/or in patients with local advanced cervical cancer whose response to CCRT is poor [17][18][19]. Performing surgery a few weeks after the end of conventional treatment offers the advantage of avoiding radiationinduced fibrosis in the pelvic tissues, which will hinder surgical dissection due to inflammation and adhesion formation resulting in an increased risk of morbidity [11]. Of the 105 surgeries performed for curative purposes, 59 (56.1%) had residual tumour with the majority beginning in complementary class 1 hysterectomies (13/17, 72.4%) who did not receive brachytherapy. Sixteen (15.2%) experienced postoperative complications, 13 belonging to the group of 83 patients previously treated with complete CCRT (15.6%) vs. 1 of 22 (4.5%) of those in the incomplete CCRT group. (p = 0.2912) The number of complications included 7/75 (9.3%) in hysterectomies with complete CCRT and 9/30 (30.0%) in exenterations of which 8/21(38.0) had complete CCRT and 1/9 (11.1%) incomplete. The lower number of complications observed in this series in patients without brachytherapy is attributed to the fact that the increased dose of radiation received in the pelvis favors the inflammation and fibrosis processes described, with a higher risk of morbidity during surgical dissection [11]. Major complications in patients treated with salvage surgeries due to persistent or recurrent CC after radiation consists of serious events that must be resolved judiciously. Serious adverse events have been reported in 16-27% of patients treated with hysterectomies after radiotherapy, the majority being class 3 hysterectomies [9,14,22]. In the series by Mazeron et al [22] with 54 cases, the figure was 25.9%; the report by Mabuchi S et al [9] with 31 patients 27%; and in Fanfani et al [14] with 73 patients, the figure was 16.4%. For PE, the numbers of serious adverse events are higher and generally reach 60% or more [20]. In the present series, 9/ of them with complete CCRT and 1 with incomplete CCRT. The analysis of the 105 patients with common variables and surgeries for curative purposes showed a median OS of 24 months. For DFS and OS, univariate analysis and multivariate analysis of OS showed significant differences in favour of age < 35 years (p=0.031, p=0.051 and p=0.007) adenocarcinoma vs squamous carcinoma (p=0.023 and p=0.019) and for the absence of lymph vascular infiltration (p=0.003 and p=0.001). The multivariate analysis for DFS and OS only showed significant differences for the absence vs presence of lymph vascular invasion. (p=0.001 and p<0.0001). For age, we selected the analysis of the groups 35 years old and under and 36 years old and greater because a previous publication on prognostic factors for pelvic exenteration for cervical cancer demonstrated that patients aged 35 years and younger had a better prognosis [20]. The differences in prognosis observed in the univariate analyses in favour of adenocarcinoma in which conventional therapy does not offer the desired results [17][18][19] suggests that the treatment modality employed by the authors could be useful in the treatment of advanced lesions of this histologic variety, although it will be necessary to have more data accumulated. Lymph vascular infiltration negatively influenced the prognosis of the 105 patients in the multivariate analyses for DFS and OS (p=0.001 and p=0.003) and has been previously reported [23,24]. For hysterectomies, the OS was 32.4 months and 38 (56.7%) patients were free of disease 25 months and greater during follow-up. The following variables were found with statistical significance in univariate analysis: age equal to or less than 35 years (p=0.004 and p=0.007) and class 3 hysterectomies (p=0.003 and p=0.003). For the multivariate analysis, statistical significance was found for age equal to or less than 35 years (p=0.004 and p=0.021) and class 3 hysterectomies (p<0.001 and p=0.003) suggesting the need for more radical surgical treatment of these patients when hysterectomies are indicated. Mabuchi et al reported [9] 30 radical hysterectomies due to persistent or recurrent CC with the 3-year survival rate being 53%. Fanfani et al [14] compared the results obtained for stage III CC in 77 patients treated with CCRT vs. 73 to whom radical hysterectomy was added to the management without ICBT. The 3-year DFS range was 62.9% for the first group and 68.3% for the second (p = 0.0686). Houvenaeghel et al [13] reported the results in 35 patients in stages IB-IVA treated with CCRT plus surgery with 26 radical hysterectomies (21 with Para aortic lymphadenectomy) and 8 pelvic exonerations. In 17 patients (48.5%), there was a complete response with a reported 10-year DFS of 66.4% of the cases in the presence of lymph node metastases (p=0.032). Also noteworthy is the absence in this series of differences for DFS (p=0.080) and OS (p=0.112) when comparing the results obtained with the use of CCRT vs tele therapy plus chemotherapy. The latter scheme was accompanied by less surgical morbidity (1/22, 4.5%, vs 13/83, 15.6%) which suggests its safe to use in patients who are unable to complete their treatment regimen [8,14]. The results obtained with pelvic exonerations showed failure of surgical treatment in 5/30 (16.6%) patients. DFS and OS for the 25 patients were 18.7 and 20.2 months respectively. The univariate analysis for DFS showed variables with statistical significance for prognosis being the absence of residual tumour in the surgical specimen (p=0.034), the report of central tumour (p=0.024) and parametrial infiltration (p=0.007). For multivariate analysis, parametrial infiltration (p=0.008 and p=0.012), for DFS central tumour (p=0.008) and rectal invasion for OS (p=0.006) were statistically significant. These results are consistent with those reported in the literature. [25,26]. regarding the benefit of adjuvant hysterectomy after concurrent chemo radiation for locally-advanced cervical cancer is controversial; the met analysis recently published by Weijia Lu et al [27]. Indicated that the recurrence rate may be higher in patients undergoing CCRT without hysterectomy. In this series, 13/105 (12.3%) patients had recurrences and in 7 (53.8%) cases the recurrences were distant, 4 of them without loco regional tumour activity. Eleven patients died with tumour activity and 2 (15.3%) patients diagnosed with adenocarcinoma obtained control of their disease for 36 and 24 months, one with local recurrence through ICBT application that she had not previously received and the other with the combination of chemotherapy agents. The observation that 53.8% of recurrences in this series were distant suggests consideration of chemotherapy for patients with lesions at risk for recurrence with reported parametrical, bladder and/ or rectal invasion, lymph vascular infiltration and lymph node metastases.

Conclusions
In this series, 83 patients received complete CCRT treatment with 13 (15.6%) having MC vs 1 of 22 (4.5%) treated with incomplete CCRT. The Kaplan Meier curves showed DFS and OS in the 105 hysterectomies and pelvic exenterations with a median follow up of 20 months and 24.0 months respectively. For complementary or salvage hysterectomies, the mean OS was 32.4 months. For pelvic exonerations, OS was 20.2 months. The analysis of the 105 cases treated with radical surgery demonstrated by univariate analysis statistical significance in favour of age < 35 years and the absence of lymph vascular infiltration. The multivariate analysis for DFS and OS only showed significant differences for the absence vs presence of lymph vascular invasion. In hysterectomies, the multivariate analysis for OS showed the following variables with statistical significance: age equal to or less than 35 years, class 3 hysterectomies and the presence of lymph node metastases. In pelvic exonerations, the univariate analysis for DFS showed variables with statistical significance for prognosis being the absence of residual tumour in the surgical specimen, the report of central tumour and parametrial infiltration. The results obtained in this series with adjuvant or rescue surgeries suggest Volume  that it could be used in patients with advanced adenocarcinoma in which the conventional treatment offers less favourable results than those obtained with squamous carcinoma. In this series, 53.8% of tumour recurrences were detected at distant sites with or without tumour in the pelvis, which suggests the benefits of using adjuvant chemotherapy in patients at risk for recurrence.