Abstract
Background/Aim: Complete clinical response in rectal cancer after neoadjuvant chemo-radiotherapy is challenging. Indeed, indication to surgery vs. “watch and wait” is a debate due the poor predictive value of restaging exams in order to identify a pathological complete response (pCR). Improving the knowledge on mutational pathways such as MAPK/ERK could be helpful in assessing the real impact of disease on prognosis and in choosing the best therapeutic target. This study aimed to evaluate the significance of biomolecular parameters as prognostic factors in patients undergoing radical surgery after chemo-radiotherapy. Patients and Methods: A retrospective analysis was performed including 39 patients who had undergone radical surgery after neoadjuvant chemo-radiotherapy for rectal adenocarcinoma stage II-III through additional evaluation of the following biomolecular markers on surgical specimens: exons 2, 3 and 4 of the KRAS and NRAS genes and exon 15 of BRAF by pyrosequencing. Kaplan–Meier survival curves were plotted to evaluate the association of pathologic response and RAS status with progression-free survival (PFS) and overall survival (OS). The log-rank test was used to assess statistical differences among the survival curves. Results: Data analysis showed RAS mutation in 15 patients (38.46%). pCR was achieved in seven patients (18%), including only two RAS mutation cases. The distribution of evaluated variables was homogeneous in the two groups based on pathological response. The Kaplan–Meier curve showed poor outcomes in OS and PFS in patients with RAS mutation (p=0.0022 and p=0.000392, respectively), but no significant differences based on pathological response for both OS and PFS. Conclusion: RAS mutation seems to be related to poor prognosis and increased risk of recurrence in rectal cancer patients undergoing radical surgery after chemo-radiotherapy.
In locally advanced rectal cancer the outcome of primary response of the tumour after neoadjuvant chemo-radiotherapy is heterogeneous. Some tumours respond completely, achieving a complete pathologic response (pCR), which is defined as the absence of tumour cells visible under the microscope after pathologic examination of a surgical specimen. This is achieved in 20-30% of cases, according to the literature, while in other cases, response is very limited (1, 2).
The role of subsequent radical surgery is widely debated in towards clinical complete response (cCR), defined as the absence of clinically detectable residual tumour. Indeed, there is a large body of literature on an observational approach commonly known as “watch and wait” (1, 2). Concomitant chemo-radiotherapy and subsequent surgery are often accompanied by a high complication rate, reduced organ function often associated with incontinence and a negative impact on quality of life (3).
At present, there is no prospective validation as to which surgical approach, whether radical or local, is best, especially when there is clinical evidence for a complete response. In fact, it has been pointed out in many studies that a cCR does not always result in a pCR, and foci of residual tumour can be identified in about 75% of surgical specimens that previously had clinical evidence for complete response (1, 4, 5).
Different studies have reported excellent outcomes in patients who achieved a pCR. A systematic review including 1,263 patients with a pCR after neoadjuvant treatment and 2,100 patients with non-complete pathologic response showed a 5-year survival of 90% and disease-free survival of 87% (5).
Rectal exploration, computed tomography (CT), positron emission tomography (PET)/CT and magnetic resonance imaging (MRI) have been shown to be relatively unreliable in predicting a pCR, with accuracy ranging from 40% to 80%. Moreover, a major problem is the risk of lymph node metastasis, despite complete local regression. In ypT0 patients treated with total mesorectal excision, the risk of lymph node or mesorectal site metastasis is 12% (6, 7).
Recent developments in oncogenes and biomarkers have raised hope as possible aid in choosing the best therapeutic strategy. The presence of mutations in the MAPK/ERK pathway (e.g., BRAF, KRAS, NRAS, ERBB2 and ERBB3) is common in rectal cancer (8, 9). The expression of EGFR levels and the activity of intracellular signalling pathway components influence the degree of pathological response to treatments with a potential predictive role. A 2018 study showed higher recurrence-free survival in patients who were wild type for the RAS/BRAF mutation, regardless of pathological response, albeit with a lack of direct correlation with response to chemotherapeutics in the neoadjuvant regimen (10).
KRAS is a proto-oncogene that seems well recognised as a predictive marker in colon cancers, although a prognostic role for it is still debated. There are data in the literature showing a higher frequency of the KRAS mutation prior to pre-operative radiotherapy in tumour tissue, compared with physiologic tissue, with higher rates in normal mucosa at 2 cm distance from the tumour margin, compared with a greater distance (4 or 6 cm). The same mutational frequency, both in pathologic and non-pathologic tissue, decreases significantly after radiotherapy (11, 12).
It also appears that a specific KRAS mutation may lead to different responsiveness to neoadjuvant combination treatment, suggesting a role for KRAS testing in predicting response to therapy. A 2013 study (13) of 60 KRAS-mutated patients defined an 8% correlation rate with pCR, compared to 13% for patients who did not express the mutation, with statistical significance (p=0.006). Notably, none of the patients with a codon 13 mutation achieved a pCR (p=0.03). More recently, Chow et al. (14) showed one case of pCR in a population of 14 patients with the G12V mutation and reported similar results in a G13D mutation cohort of 15 patients (one patient achieved pCR). In contrast, eight of 41 patients with the G12D mutation reached a pCR (20%).
The aim of the study was to identify the extent at which evaluation of biomolecular parameters such as KRAS, NRAS and BRAF can impact the prognostic outcome in rectal cancer patients eligible for neoadjuvant chemoradiotherapy. The secondary aims were to observe how biomolecular parameters impact cCR subpopulation and assess their potential impact on the “watch and wait” approach.
Patients and Methods
We performed a retrospective analysis of 39 patients who had undergone surgery for rectal adenocarcinoma from May 2011 to June 2017 at the Guglielmo da Saliceto Hospital, Piacenza, with stage II (T3-4, N0) or III (each T, N1-2), according to American Joint Committee on Cancer (AJCC) guidelines, with a distal tumour margin within 15 cm of the anal rim by rigid coloscopy or NMRI (15). Local staging was performed by rectal endoscopy or MRI. To exclude metastatic pathology, chest abdomen CT with staging contrast medium was performed. If adequate pre-operative tissue specimen collection was not possible, biomolecular assessments were performed on the surgical specimen. pCR was defined as the absence of detectable tumour in both the entire surgical resection and the regional lymph nodes.
All patients received fractionated neoadjuvant radiation treatment at 1.8 Gy per day, five fractions per week, for a total tumour dose of 45 to 50.4 Gy. Twenty-seven patients were treated with concomitant capecitabine at a dose of 825 mg/m2 die for the duration of radiation therapy (16), and 10 patients received concomitant 5-fluorouracil continuous intravenous infusion (225 mg/m2/day) (13). Surgery was scheduled 6-8 weeks after the completion of pre-operative treatment. All patients underwent anterior resection of the rectum, with the exception of two patients who underwent abdomino-perineal amputation surgery (Miles procedure), one patient who underwent low resection and one patient who underwent ultra-low resection of the rectum. Twelve patients subsequently underwent adjuvant chemotherapy for a total of 6 months by stage of disease and histological parameters, according to national guidelines; of these, 10 patients were treated according to the XELOX scheme (capecitabine 1,000 mg/m2 on day 1 to 14 every 3 weeks, oxaliplatin 130 mg/m2 on day 1 every 3 weeks) (17), and only one patient was treated according to the FOLFOX scheme (oxaliplatin 85 mg/m2 on day 1 every 2 weeks, leucovorin continuous infusion 200 mg/m2 on day 1 every 2 weeks, 5-fluorouracil bolus 400 mg/m2 on day 1 every 2 weeks, 5-fluorouracil 600 mg/m2 continuous infusion of 46 hours on day 1 every 2 weeks) (18).
Follow-up. The patient follow-up protocol consisted of medical evaluation with rectal exploration, evaluation of blood tests including complete blood count, hepatorenal function, CEA and CA19.9 markers every 6 months for the first 5 years. A CT scan of the chest, abdomen and pelvis with contrast medium was performed every year for 5 years; each of these scans was followed 6 months later by an ultrasound of the complete abdomen and pelvis. An MRI of the pelvis with contrast medium was also considered, based on the clinician’s judgment. Colonoscopy was performed 1 year after surgery, after 3 years and then every 5 years thereafter, according to current international guidelines (19). OS was recorded at every scheduled oncological visit and was included in the survival analysis only for related rectal cancer exitus.
DNA extraction. Histological specimens were fixed in 4% buffered formalin and then included in kerosene. A pathologist evaluated the stained sections with haematoxylin-eosin, identifying areas of increased presence of tumour cells and attempting to eliminate, as much as possible, any necrosis and healthy tissue. Each sample was macro-dissected to process at least 200 tumour cells with an enrichment greater than 50 percent (neoplastic cell/healthy cell ratio) to ensure the sensitivity and reliability of the biomolecular assay. DNA was extracted by treatment with lysis buffer and proteinase K at 58°C and then stored at 4°C.
Tumour regression grade evaluation. Tumour regression grade (TRG) evaluation was performed by microscopic analysis by an anatomic pathologist using the seventh edition of the AJCC Cancer Staging Manual, a four-point system modified from Ryan & Coll.: A score of 0 indicates a pathologic complete response (i.e., no tumour cells visible under the microscope), whereas a score of 1 indicates a ‘near’ complete response (i.e., the presence of single tumour cells or rare small aggregates), a score of 2 identifies a tumour remnant with obvious signs of regression but with the presence of a partial response, and a score of 3 indicates an extensive tumour remnant with no obvious signs of regression (little or no disease response). Note that metastatic lymph nodes are not included in this evaluation.
Mutational analysis. Exons 2, 3 and 4 of the KRAS and NRAS genes and exon 15 of BRAF were analysed by pyrosequencing. Diatech Pharmacogenetics Anti-EGFR MoAb Response® kits (KRAS status and NRAS status) identified the major variants of codons 12, 13, 59, 61, 117 and 146 of the KRAS gene and codons 12, 13, 58, 59, 61, 117 and 146 of the NRAS gene. Regarding the BRAF gene, the Anti-EGFR MoAb Response® kit (BRAF status) evaluated variants of codon 600 of exon 15.
Statistical analysis. Patients were registered with a unique recognition code in a Microsoft Excel file (Microsoft Office version 2010). Quantitative variables were described by mean±standard deviation, and qualitative variables were described by absolute and percentage frequencies. Normality was checked for all continuous variables. Comparisons of covariates were conducted using Pearson’s χ2 test or Fisher’s exact test for categorical variables and a t-test for continuous variables. PFS and OS were estimated by the Kaplan–Meier method and compared by the log-rank test. All analyses were performed using RStudio version 3.6.0 statistical software with two-sided significance tests and a 5% significance level.
Results
Data analysis showed the presence of KRAS mutation in 14 patients (35.9% of the total), including one in codon 146T, one K117N, two G12V, five G12D and five G13D. Only one patient (2.6%) showed an NRAS G12D mutation (Table I).
Patient characteristics.
The distribution of the evaluated variables, such as sex, age at diagnosis, comorbidity, Eastern Cooperative Oncology Group performance status (ECOG-PS), anatomopathological features (grading, vascular invasion and mucinous histotype), stage of disease (II or III), neoadjuvant treatment, post-surgical morbidity and RAS mutations, were homogeneous in the two groups on the basis of the pathological response, with the exception of TRG and adjuvant treatment (none of the patients with pathologic complete response received adjuvant therapy) (Table I).
Of the 39 patients, only seven achieved a pCR (18% of the total population). Among them, two patients (13% of the total) had a KRAS mutation (in one case, G13D; in the other, K117N); five were wild-type patients (21% of this subgroup of patients). The Kaplan–Meier survival curves revealed no significative difference in both OS and PFS among the pathologic response (Figure 1 and Figure 2; p=0.11 and p=0.47, respectively), although worse outcomes were observed for non-pCR patients.
Kaplan-Meier curve of 5-year OS based on pathological response. The analysis included patients who reached a pCR vs. patients who did not reach pCR. OS: Overall survival; pCR: pathological complete response.
Kaplan-Meier curve of five-year PFS based on pathological response. The analysis included patients who have reached the pCR vs. patients who have not reached pCR. PFS: Progression free survival; pCR: pathological complete response.
The 5-year OS was significantly lower in RAS-mutated patients than in RAS wild-type patients (p=0.0022) (Figure 3), and the cause of death was related to disease progression in eight of 10 patients. Exitus occurred in 10 out of 39 patients (25.6%), including seven with RAS mutation (three patients with G12D mutation, three with G13D mutation and one with G12V mutation). None of these had reached a pCR after neo-adjuvant therapy. The 5-year PFS in the Kaplan–Meier curves was also significantly different between RAS wild type and mutated patients (p=0.00039) (Figure 4). Disease recurrence occurred in nine patients (23.07%) (four with lung progression, two with hepatic and peritoneal progression, one with hepatic recurrence, one with local and bone recurrence and one with perineal recurrence). Mutations were documented in eight out of nine patients in this group (three expressed a G13D mutation, three a G12D mutation, one a G12V mutation and one a 146T KRAS mutation) (Table II). Going on to assess the stage variation in response to neoadjuvant treatment, 26 patients (66.7%) achieved disease regression, while four (10.3%) underwent progression despite treatment (refractory patients). Ten patients (25.6%) of the total sample underwent regression, and only one developed tumoral relapse during follow-up (2.6%).
Kaplan-Meier curve of five-year OS based on RAS status. The analysis included patients with the presence of RAS mutation vs. patients of the wild type. OS: Overall survival; RAS MUT: RAS mutation; RAS WT: RAS wild type.
Kaplan-Meier curve of five-year PFS based on RAS status. The analysis included patients with the presence of RAS mutation vs. patients of the wild type. PFS: Progression free survival.
Recurrence of primary rectal cancer (including local and distance recurrence) according to characteristics the patient population.
In a secondary analysis of mutated patients, those who did not respond to neoadjuvant treatment mainly belonged to the G12D (three out of five patients), G13D (one unchanged patient and one refractory patient progressed after neoadjuvant therapy; thus, two out of five patients) and G12V (only one out of two patients with such mutation) mutational categories. Notably, the percentage of patients with RAS mutation increased in line with an increase in the TRG value (55.6% for TRG 3 patients); however, the value was not statistically significant (p=0.57) (Figure 5).
Frequency of RAS mutations according to the four grades of TRG classes: TRG 0: pathologic complete response (i.e., no tumour cells visible under the microscope); TRG 1: ‘near’ complete response (i.e., the presence of single tumour cells or rare small aggregates); TRG 2: identifies a tumour remnant with obvious signs of regression but with the presence of a partial response; TRG: 3 indicates an extensive tumour remnant with no obvious signs of regression (little or no disease response). TRG: Tumour regression grade.
Post-operative surgical morbidity within 90 days was detected in six patients (15.4%), and only one case was intraoperative, due to concomitant compartment syndrome of the lower legs in the KRAS G12D mutated patient. Other reported morbidity included two anastomotic leaks and one case of enteric fistula that underwent surgical re-laparotomy, while in other cases, non-surgical management was achieved for sub-occlusion and acute renal injury due massive output from loop-ileostomy. Delayed morbidity within 24 months was reported in two cases: one patient with chronic sinus of colorectal anastomosis with stenosis and one relaparotomy due mechanic ileus for peritoneal adhesion after 12 months from surgery.
Discussion
According to primary aim of this study the evaluation of KRAS could be useful to assess the prognosis in patients underwent neoadjuvant chemo-radiotherapy in rectal cancer. Currently, available literature shows that KRAS mutation is an independent factor associated to worse prognosis and in addition KRAS mutation is related to advanced pathologic T stage, tumor deposits, perineural invasion and elevated carcinoembryonic antigenic levels (20). In addition, a meta-analysis showed that KRAS mutation is an indicator of PFS and not related to pCR (21). Further, the multicentre RASCAL (22) study evaluated the prognostic significance of codon 12 or 13 mutations in 2721 patients from 13 states; in multivariate analysis, only the mutation at codon 12 was independently associated with an increased risk of recurrence and death.
In fact, our prognostic outcomes OS and PFS were negatively affected by evidence of RAS mutation, with survival curves separating statistically significantly (p=0.0022 and p=0.00039, respectively). Of the patients with RAS mutation, only two achieved a pCR (13.3%). This result seems to coincide with those of other studies (12), in which the cohort of mutated patients had pCR rates from 14% to 15%, while in wild-type patients, the percentage increased from 33% to 34%.
In our analysis, achieving a pCR was not statistically significantly associated with improved OS and PFS, although the survival curves had a better trend in patients who achieved a pCR. These data, although they may be related to a low sample size, did not contrast with existing literature in supporting the use of pCR as an appropriate surrogate endpoint for 5-year OS in patients with rectal cancer treated with neoadjuvant therapy (23).
In descriptive analysis, the two groups pCR and non-pCR were homogeneous in terms of the variables of interest, except for adjuvant therapy and TRG (Table I). The former finding is justified by the fact that following relief of a pCR, the clinical choice was unbalanced in favour of instrumental clinical follow-up over adjuvant therapy. Regarding the TRG figure, as might be expected, there was an imbalance in favour of TRG 0, as it corresponds to pCR.
Going on to better characterise the mutated populations that seem to respond less to neoadjuvant treatment, the population with G13D mutation seemed to be the one with the most unfavourable features (of which one patient even progressed during therapy). In fact, out of three mutated patients who had lymph node metastases in the surgical report, two had a G13D mutation. This finding confirmed the association of the specific mutation with a higher risk of death, highlighted by the CRYSTAL and OPUS studies in adjuvant treatment (24, 25).
A total of seven patients in the anatomopathological report after surgery had lymph node metastasis; of these patients three had the KRAS mutation (42.9%), corresponding to 20% of mutated patients, compared to 16% of wild-type patients; among the mutated patients, two expressed a G13D mutation and only one a G12D mutation. Our data were in line with existing literature showing a correlation between KRAS mutation and nodal spread (26).
In addition, the percentage of patients with RAS mutation who did not respond to neoadjuvant treatment (33.3%) was considerable. This finding confirms the trend found in larger studies involving rectal cancer patients (20, 21). In fact, in this population the percentage of KRAS mutations appears to be constant between 36% and 37%, confirming the possibility of generalizing the possible prognostic significance of the finding.
Within the limitations of the low sample size, the mutational categories G12D (three out of five patients), G13D (one unchanged patient and one refractory patient who progressed after neoadjuvant therapy, i.e., two out of five patients) and G12V (only one out of two patients with such mutation) seemed to be the most aggressive.
In an overall comparison of all known codon 12 mutations with codon 13 mutations, it appears that the former correlated unfavourably predictively in response to drug therapy. In fact, four patients out of a total of six who did not go on to regression after chemotherapy had a codon 12 mutation. Laboratory results seem to support that codon 12 mutation of KRAS, compared with codon 13 mutation, manifests an increased ability for cell transformation, increased growth independent of cell anchoring and increased ability to suppress apoptosis (27, 28).
Divergent transduction signals between KRAS codon 12 and 13 mutations and an association between activation and attenuation of various pathways, including AKT, JNK and FAK, and resistance to many chemotherapeutic agents and radiation treatment in multiple tumours have been demonstrated in in vitro studies (29, 30). Thus, it seems that KRAS may have a role as a negative prognostic factor after neoadjuvant chemoradiotherapy treatment, even in the case of potential cCR (28). Indeed, an absence of lymph node metastasis on pre-operative imaging cannot be accurately guaranteed, which is a huge problem because the risk of lymph node metastasis, despite complete local regression, is frequent. KRAS could play a major role in such prediction, given the high prevalence of lymph node metastasis in patients carrying the mutation, especially in codon 12 and G13D (26). In this subgroup, more aggressive treatment with surgical resection could be the best option, even in the presence of radiologic and endoscopic evidence for cCR. Nevertheless, this finding requires confirmation in an extensive population study.
As this study is limited by low population size, the results are still partial; nevertheless, they can still be considered marginally indicative. Other study limitations are related to treatment heterogeneity, given the retrospective data; the population is representative of patients with rectal cancer who were candidates for neoadjuvant therapy, according to guidelines, as judged by the clinician. Treatment heterogeneity may predictably have affected response rates or directly impacted the association of mutational status with pCR data. Adjuvant chemotherapy treatment was sometimes added after surgery, depending on the extent and histologic parameters, according to the guidelines available in the study period, either as monotherapy or with a 5-fluorouracil-containing doublet (FOLFOX4, XELOX).
Increased knowledge in the field of molecular biology seems to be helping oncologists to choose the best therapeutic strategy; in the current literature more and more biomarkers are being studied as prognostic or predictive factors and the expression of other regulatory factors such as miRNA was investigated in rectal cancer treated for local recurrence: of these only miRNA-21 showed a difference in expression between healthy and malignant tissue (31).
Biomolecular assessment of RAS (NRAS and KRAS) is now usable in current clinical practice, however, at present, international guidelines do not include the use of biomolecular determinations in the non-metastatic setting (19). Since nationwide usability is high, and mutational evaluation might be useful at a later stage if there is a distant relapse of disease, it would seem possible and desirable to also use this parameter because of its prognostic and predictive role of response to neoadjuvant therapy, although such therapy does not currently use driver monoclonal antibodies. In particular, the medical-scientific community is increasingly considering the possibility for patients who achieve a complete clinical response to imaging and endoscopy examinations to take a “watch and wait” approach, thereby sparing surgery at first instance. Such an approach could be resorted to at a later stage, upon evidence of local recurrence of disease. This strategy is based on evidence that patients who go on to a complete pathologic response are those with a favourable prognosis; they incur disease recurrence more rarely and survive longer. For such patients, an observational approach could be considered; however, a complete clinical response does not always coincide with a complete pathologic response. At the moment, no strengthening factors are available to predict pCR in order to help the clinician opt for the “watch and wait” approach with respect to surgery. RAS evaluation seems to be able to aid the clinician’s choice as an unfavourable prognostic factor in the presence of mutation. The data from our case series seem to confirm this assumption. In the face of a different degree of tumour regression, patients with RAS mutation incur disease recurrence more easily and have a higher disease-related mortality than wild-type patients. TRG assessment, a known prognostic factor after neoadjuvant therapy, also seems to associate in part with the RAS profile, expressing lower tumour regression in patients with RAS mutation.
In conclusion, the presence of a RAS mutation at both NRAS and KRAS levels could be a risk factor for poor prognosis for cCR as well, despite the low number of NRAS mutations in our population.
Future research is necessary to confirm RAS as a risk factor in non-metastatic rectal cancer, and, in particular, its use in a decision-making algorithm for the “watch and wait” strategy.
Footnotes
Authors’ Contributions
Elena Orlandi designed the study, analysed the data and wrote the manuscript. Elena Orlandi and Luigi Cavanna conceptualised the report. Andrea Romboli, Mariangela Palladino, Stefano Vecchia and Serena Trubini collected the patients’ clinical data. Chiara Citterio performed the statistical analysis. All Authors revised the final version of this report.
Conflicts of Interest
The Authors declare that there are no conflicts of interest.
- Received January 9, 2023.
- Revision received January 31, 2023.
- Accepted February 21, 2023.
- Copyright © 2023 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.