Introduction

Up to 70 % of patients with non-metastatic rectal cancer present with locoregionally advanced disease [1]. Locally advanced rectal cancer (LARC) is generally defined as stage II and III rectal cancer. Neoadjuvant chemoradiotherapy (CRT) followed by a total mesorectal excision (TME) remains the globally accepted method for the management of rectal LARC [2].

With TME, the incidence of morbidity ranges from 6 to 35 %, which includes anastomotic leaks, blood loss, and sexual dysfunction resulting from the procedure. In addition, the mortality rate reaches up to 2 % [3]. The length of hospital admissions ranges from 8 to 15 days [4–6]. There are also compelling data regarding the effect of resections on patients’ quality of life; deterioration in bowel function is common following anterior resection, and patients with low-lying cancers may require a permanent stoma, which would be associated with psychological morbidity [7].

Neoadjuvant CRT has become the standard treatment for patients with LARC, allowing for a reduction in local recurrence and an increase in sphincter preservation [8, 9]. Neoadjuvant CRT followed by surgery 6–8 weeks later may result in a complete pathologic response (cPR) in up to 15–30 % of cases [10], and it is also associated with a better oncologic outcome.

New trends have suggested the possibility of neglecting planned surgical resections after neoadjuvant treatment in cases of extensive tumor response. In addition, more studies have addressed the use of a non-operative “watch and wait” policy in patients with a complete clinical response (cCR) [8, 9]. There is growing evidence showing that regimented clinical assessment after CRT can identify patients with cCR, allowing for the avoidance of immediate radical surgery while preserving good oncologic outcomes.

To evaluate the benefit of a watch and wait policy in patients with cCR treated with pre-operative CRT, we compared the oncologic results of a group of patients with cCR followed with a watch and wait policy versus those followed with radical surgery at our institute.

Materials and methods

Study inclusion and design

The study group consisted of a consecutive series of patients treated at Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan, Republic of China, from 2007 to 2014 for primary rectal cancer (up to 10 cm from the anal verge). All the patients had biopsy-verified rectal adenocarcinoma. This study was approved by the Institutional Review Board of the Tri-Service General Hospital (TSGHIRB No.: 2-103-05-076). Informed consent was obtained from all the patients.

Pre-treatment oncological staging comprised abdominal imaging (computed tomography [CT] or magnetic resonance imaging [MRI]), and tumor marker analysis (carcinoembryonic antigen [CEA] and cancer antigen [CA] 19–9 levels). As a general rule, the highest stage for each parameter evaluated (T, N, circumferential margin, involvement of adjacent organ, M) was considered to be the definitive pre-treatment stage [11, 12].

Patients selected for neoadjuvant CRT were required to meet all the following criteria: (a) a biopsy-proven rectal adenocarcinoma, (b) a tumor location up to 10 cm from the anal verge. (c) a primary clinical stage of II–III, and (d) an Eastern Cooperative Oncology Group performance status of 0–2 [2].

5-Fluorouracil (5-FU), as a single drug or in combination with other drugs (leucovorin, carboplatin, or oxaliplatin), was administered by bolus or continuous venous infusion. During this study, a single standard regimen was used; initially, 5-FU was administered as a bolus (350 mg/m2/day) with a low-dose leucovorin bolus (10 mg/m2/day) for 5 days on days 1–5 and 29–33 in combination with radiotherapy (45 Gy in 25 fractions or 54 Gy in 30 fractions) [13].

Patients were assessed for their tumor response 8–12 weeks after the completion of chemoradiotherapy with the same clinical and radiologic tools used in the baseline assessment of the tumor extent. All patients who were considered clinical complete responders according to rigorous criteria of clinical, endoscopic, and radiologic findings were treated without immediate radical surgery. The three criteria for cCR were (a) the absence of a residual ulceration, mass, or mucosal irregularity upon clinical/endoscopic assessment; (b) whitening of the mucosa and the presence of neovasculature (telangiectasia); and (c) radiologic imaging, such as CT, transrectal ultrasonography (TRUS), or MRI, without evidence of extrarectal residual disease, which were necessary for patients to be considered to have cCR [14]. For example, lymph nodes <5 mm in their transverse diameter, or absence of irregular borders were considered negative for metastases, on standard T2-weighted MRI study or on CT scans. In addition, there was no evidence of hypoechoic, inhomogeneous lesion with irregular borders; thickening or destruction of the bowel wall under TRUS was also considered negative for extrarectal disease. The presence of clinical or endoscopic features of an incomplete response to CRT and the radiologic evidence of residual disease within the mesorectum were diagnostic of an incomplete clinical response, and therefore, radical surgery was recommended.

Experienced pathologists handled the pathologic evaluation of resected rectum specimen. The evaluation including (a) recording the length and the diameter of the rectum; (b) looking for and documenting the presence and appearance of any visible or palpable lesion and its location relative to the margins and to any landmarks; (c) dissecting the mesentery and thin section of the mesenteric fat, and examining and palpating each section for lymph nodes; and (d) submitting for histology the entire tumor site, each identified lesion, and lymph node. The percentage of pathological response ranged from no evidence of treatment effect (0 %) to complete regression with no viable tumor cells identified (100 %) and ypT0, corresponding to a percentage of pathological response of 100 %.

We identified 829 patients with biopsy-proven rectal adenocarcinoma. Next, medical records were queried for primary clinical stage, complete CRT therapy, and surgery status; 562 patients were excluded because of stage I and IV disease (n = 200, and 137, respectively), failed in finishing the CRT (n = 103), and surgery without CRT (n = 122). Thus, 267 patients remained.

Of these, 223 patients were also excluded because of an incomplete clinical response. Only 44 patients with LARC undergoing complete CRT with cCR at our institute were included in the study (Fig. 1). In all cases, the decision to employ a watch and wait policy or to perform radical surgery was made after thorough discussions between the surgeon and the patient. Patients in the watch and wait group were informed that disease recurrence might occur at any moment during the follow-up period. If positive nodes were identified, radical surgery would be advised. Local recurrence was defined as the presence of adenocarcinoma within the rectal wall, with negative nodes, and no distal metastases on imaging. Patients with local recurrence were referred for transanal wide excision.

Fig. 1
figure 1

Flowchart of patient selection

Full size image

All patients with cCR were evaluated with outpatient visits every 3 months by a single experienced colorectal surgeon who performed a clinical examination in addition to a rigid proctoscopy or colonoscopy (Fig. 2). Biopsy was utilized selectively on any residual nodularity or scarring on exam. CEA levels were measured at the time of restaging (after finishing CRT at 8–12 weeks) and every 3 months thereafter. At the third year of follow-up, patients were examined every 6 months. The CEA cutoff value is ≤5 ng/dl at our institute, which has been demonstrated to have significant prognostic value in some studies [15]. A radiologic imaging modality (including chest radiography, CT scans, and MRI) was used to exclude mesorectal disease and systemic status after 6 months and annually thereafter. CT scans were routinely obtained for all patients.

Fig. 2
figure 2

Colonoscopic findings in the patient with complete clinical response. a, b The lesions pre- and post-chemoradiotherapy from one patient respectively. c, d The lesions pre- and post-chemoradiotherapy from another patient respectively. a, c One huge reddish polypoid mass occupying the lumen of the colon. b, d A whitening lesion without ulceration or mucosal irregularity and presence of neovasculature

Full size image

Statistical analysis

For each patient, CEA levels were evaluated at diagnosis, after CRT, and every 3 months during follow-up. The mean CEA levels are presented in Table 1. Univariate associations between categorical variables were assessed using Fisher’s exact test. The 5-year overall survival and disease-free survival rates were calculated with the Kaplan–Meier method (Fig. 3). The statistical analyses were performed using SPSS software (ver. 15.0, SPSS Inc., Chicago, IL, USA).

Table 1 Demographic and clinical characteristics of the study group
Full size table
Fig. 3
figure 3

a Disease-free survival and b overall survival for patients with a clinical complete response following “watch and wait” policy and patients following “radical surgery”

Full size image

Results

A total of 44 patients with LARC underwent CRT with a cCR during the study period. We identified 18 patients who were treated with active surveillance, classified as the watch and wait group, and 26 patients who underwent TME, known as the radical surgery group. Of these 26 patients, 17 underwent lower anterior resection (LAR), 6 underwent LAR with diverting loop stoma, and 3 received abdominoperineal resection (APR). The watch and wait group tended to be slightly older than the radical surgery group (67.58 vs. 63.78 years, respectively, p = 0.792). The majority of patients included in this study were men (men, 61.36 %, 27/44; and women, 38.64 %, 17/44), and of the male patients, 55.55 % (15/27) were included in the watch and wait group, whereas only 17.64 % (3/17) of the female patients were in the watch and wait group (p = 0.26). The distance of the tumor from the dentate line was 3.55 and 4.61 cm in the watch and wait and radical surgery groups, respectively. The demographic and clinical characteristics of the study group are reported in Table 1.

The mean follow-up time was 49.92 and 42.28 months, for the watch and wait and radical surgery groups, respectively. There were no major complications documented related to the watch and wait group, while 13 patients developed post-operative complications in the radical surgery group. Three patients had complications related to their diverting stoma; one developed hypovolemic shock requiring fluid resuscitation, while two parastomal hernias required elective repair. Five patients developed surgical site infections, and one had an intra-abdominal abscess requiring percutaneous drainage. One patient had an anastomotic stricture solved with a diverting stoma, and three patients were documented with poor functional outcomes related to LAR syndrome.

During the follow-up for the watch and wait group, the disease-free survival time at the median follow-up was 69.78 months, and the 5-year overall survival rate was 100 % (Fig. 3). No distant metastasis was observed; local recurrence occurred in two patients, which was treated with transanal wide excision with no associated post-operative complications at 14 and 36 months, and the pathology reports were restricted to ypT1. For each patient, CEA levels were evaluated at diagnosis, after CRT, and every 3 months during the follow-up period (Table 1).

The disease-free survival time of the radical surgery group was 89.04 months, and the 5-year overall survival rate was 92.30 %. Twenty-five patients in this group reached cPR after radical surgery with TME; one patient, whose primary stage was cT3N2, presented with ypT0N1 after radical surgery. No local recurrence was found. One patient developed a single pulmonary metastasis that was treated with wedge resection at 14 months after radical surgery; this patient currently has no evidence of disease during follow-up. Two patients died of unrelated causes at 36 and 60 months after surgery, respectively; one had cardiovascular disease and one had an intracranial hemorrhage. There was no difference in disease-free survival (p = 0.354) or 5-year survival rate (p = 0.403) between the groups (Fig. 3).

Discussion

Compared with patients with cCR treated with radical surgery with TME, those treated with a watch and wait policy benefit from avoiding the immediate and long-term morbidity associated with a major operation. While patients may complain about the strict surveillance schedule, the oncologic outcome was significant at our institute.

In our study, the watch and wait group showed a much lower number of rectal tumors compared to those treated with radical surgery. We presumed that radical surgery for distal rectal cancer was APR with a permanent stoma, and many patients in the watch and wait group showed a strong desire to avoid a permanent stoma. In the radical surgery group, APR was performed on only 11.53 % (3/26) of patients, and the rest had LAR with/without a diverting stoma 88.47 % (23/26). We assumed that patients with cCR after CRT were more likely to refuse a permanent stoma, instead opting for the watch and wait policy, especially those who could not maintain bowel continuity after a curative resection such as APR. Our data also showed that radical surgery with TME carries a significant risk of post-operative complications. The results of our study should encourage further investigations to standardize the criteria of cCR, which allow practitioners the ability to more reliably identify patients who are candidates for the watch and wait policy. If diagnostic and treatment strategies can be standardized, then perhaps future patients, such as the cPR patients in our study, can avoid the morbidity associated with radical surgery while preserving optimal long-term oncologic outcomes.

The treatment of patients with cCR after CRT remains controversial. Previous clinical trials have also demonstrated that patients with cCR after CRT have both better oncological outcomes and lower rates of mesorectal lymph node metastases [12]; a watch and wait policy and organ-sparing strategies have been advocated in such patients [16–18]. The use of alternative treatment strategies without TME is desirable to avoid the significant post-operative morbidity, unnecessary enterostoma, and negative outcome. A watch and wait policy avoids post-operative complications and minimizes the risk of adverse functional outcomes in patients with cCR [16–18]. However, a watch and wait policy after CRT requires that cCR be accurately identified using patients’ clinical and radiologic features.

The clinical assessment of tumor response is a major concern for individualized patient management. Patients would be viewed as having a cCR if no visible or palpable irregularity nodules were found clinically. Habr-Gama et al. have detailed the clinical and endoscopic findings of patients with cCR; their findings include the whitening of the mucosa in the rectum, any telangiectasia, and a subtle loss of pliability of the rectal wall harboring the scar. cCR has also been described as the absence of positive signs of residual disease. An incomplete clinical response is considered in the presence of a deep ulceration with or without a necrotic ulceration, a palpable nodule, even with complete mucosal integrity, or any significant stenosis [14].

Current imaging techniques (TRUS, CT, and MRI) have been reported to be far less accurate while restaging rectal cancer after CRT. [8, 19] A CRT course may extensively modify cancer tissue and the surrounding structures, including overgrowth fibrosis, wall thickness, muscle disarrangement, tumor necrosis, calcification, and inflammatory infiltration [13].

Interestingly, Pomerri et al. also produced generally disappointing findings for the positive predictive value of lymph node status on TRUS, CT, and MRI, but the negative predictive value (NPV) was significant (TRUS, 95 %; CT, 93 %; MRI, 95 %) [20], indicating that it may be useful for the prediction of LN-negative status. It also provides strong evidence for the classification of negative lymph node status in our study; however, T status is much easier to evaluate via digital examination and colonoscopy with/without biopsy.

To avoid unnecessary surgery for LARC with cCR after CRT, the clinical assessment of post-CRT staging should be optimized. The diagnosis in our study was standardized, and patients were generally required to meet all the following criteria to be considered to have cCR: (a) the absence of a residual ulceration, mass, or mucosal irregularity during a digital examination and colonoscopy assessment; (b) whitening of the mucosa and the presence of neovasculature; and (c) radiologic imaging (CT, TRUS, or MRI) that showed no evidence of extrarectal residual disease. In our study, three experienced attending colorectal surgeons assessed clinical responses to pre-treatment therapy in each patient. For this reason, these subsets of patients were strictly adhering to the criteria described above; those with incomplete tumor regression were excluded.

Our study group showed an overall initial cCR rate of 16.48 % (44/267), which was similar to the rates of 15–30 % reported in some previous studies [10, 21]. In the radical surgery group, 25 patients showed cPR in the final pathologic reports; however, one patient presented with ypT0N1, with 1 positive in the 13 retrieved lymph nodes, and a primary tumor that was classified as cT3N2M0, stage IIIC. The concordance rate was 96.15 % (25/26) for initial cCR to be cPR, which was similar to the result for NPV under TRUS, CT, and MRI [20]. We assumed that patients whose primary tumor was staged as N ≥ 2 with a rapid response to CRT should receive two different imaging modalities to exclude the presence of micro-metastases. We supposed that this might be the key to accurately predicting cCR while using imaging modalities with the advantage of the NPV for determining lymph node status.

Follow-up included outpatient visits every 3 months by three experienced colorectal surgeons consisting of a clinical and digital rectal examination in addition to a rigid proctoscopy or colonoscopy. A radiologic imaging modality (including CT scans, MRI, and TRUS) was used to exclude mesorectal disease and to evaluate the systemic status after 6 months and annually thereafter. CEA levels were obtained every 3 months. After 2 years of follow-up, patients were examined every 6 months. Although several molecular tumor markers have been described over the past years, only CEA remains clinically significant for staging colorectal cancer [22]. The role of CEA in determining prognosis for colorectal cancer has been well documented, not only in patients with locally advanced disease but also in those with metastatic disease [15]. One study suggested that patients with a low CEA level after CRT are more likely to achieve a cCR and have a better outcome [16]. Previous studies showed that neoadjuvant CRT may provide significant local tumor control, as reflected by the significant downstaging and cancer cell necrosis resulting from increasing doses of radiation (and possibly chemotherapy). Therefore, the post-CRT CEA status could reflect the effectiveness of neoadjuvant CRT. [15] In our data, all patients presented with low CEA levels <5 ng/dl at their post-CRT follow-up (including post-CRT and 0.5, 1, 1.5, 2, 2.5, and 3 years of follow-up time, Table 1).

In the present study, the disease-free survival time of the watch and wait and radical surgery groups was 69.28 and 89.04 months, respectively, and the 5-year overall survival rates were 100 and 92.30 %, respectively. Compared to previous studies, we had similar disease-free survival and 5-year survival rates in the watch and wait and radical surgery groups. Two local recurrences were found in the watch and wait group; no local recurrence and one distant pulmonary metastasis were noted in radical surgery group. Previous studies [14, 16] had a higher rate of local recurrence than that observed in our study. This might be attributable to the differences in subjective classification of cCR between individual surgeons at each institution. If our criteria for cCR tended to be more rigorous, it could lead to lower rates of local recurrence. Radhika et al. [23] also showed significant oncological outcomes in a watch and wait group, but their data population did not completely fulfill the criteria of LARC, and patients with pre-treatment TNM stage I and IV were also enrolled into the study; in addition, they did not describe the cPR group with initial cCR. This is the first retrospective study for patients with initial cCR comparing the oncologic outcomes between watch and wait and radical surgery groups.

There were some limitations in our study. First, our approach was primarily retrospective, with a limited sample size and with patients from a single institution. Second, tumor recurrences should be considered in patients with LARC with cCR under the watch and wait policy. Habr-Gama et al. demonstrated that local recurrence may develop in 31 % of patients with an initial cCR and that more than half of these recurrences develop within 12 months of follow-up. Salvage therapy is possible in 90 % of recurrences, resulting in 94 % local disease control and 78 % organ preservation [9]. Third, during the 7 years of the study, the imaging modalities to stage rectal cancers have improved radically, while the ability to deliver radiation therapy has also evolved dramatically. Therefore, different protocols of CRT and pre-operative staging have been used over time. Fourth, a notable weakness is that patient selection into the watch and wait or radical surgery group was not randomized. The decision to treat with a watch and wait policy was mostly driven by patients’ preference for the avoidance of radical surgery with TME and permanent stoma creation. Moreover, low post-CRT CEA status may be associated with increasing rates of cCR. However, all the 44 patients (100 %) with cCR in this group had low pre-CRT CEA levels, but we could not confirm that pre-CRT CEA levels were a significant predictor of cCR.

Conclusion

This study may suggest the validity of a watch and wait policy in patients with LARC with cCR after CRT. Meticulous follow-up may play an important role in cases of cCR after CRT. In addition, a low post-CRT CEA level was strongly related to cCR in this study. If local recurrence is present after CRT, transanal wide excision may be performed as salvage treatment. From our limited experience, the patients who experienced the watch and wait policy in the cCR group may have benefitted from avoiding radical surgery, and they retained significant oncologic outcomes in terms of disease-free survival and overall survival. A watch and wait policy is a safe option for LARC after CRT with cCR. Our data also suggested that the validation of NPV with current imaging modalities might be the key to be precisely predicting cCR. Further prospective and multi-institutional studies are necessary to confirm the safety of the watch and wait policy and to create evidence-based guidelines for surgeons for the treatment of patients with cCR.