Advertisement

What are the problems with the current staging of discontinuous tumour nodules (DTNs) in colorectal carcinoma? Is there a better way?

  • Ella Karbanowicz
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, Sydney, NSW, Australia

    Sydney Medical School, University of Sydney, Sydney, NSW, Australia
    Search for articles by this author
  • Talia L. Fuchs
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    Sydney Medical School, University of Sydney, Sydney, NSW, Australia

    Douglass Hanly Moir Pathology, Macquarie Park, NSW, Australia
    Search for articles by this author
  • Angela Chou
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, Sydney, NSW, Australia

    Sydney Medical School, University of Sydney, Sydney, NSW, Australia
    Search for articles by this author
  • Loretta Sioson
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, Sydney, NSW, Australia
    Search for articles by this author
  • Amy Sheen
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, Sydney, NSW, Australia
    Search for articles by this author
  • Mahsa S. Ahadi
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, Sydney, NSW, Australia

    Sydney Medical School, University of Sydney, Sydney, NSW, Australia
    Search for articles by this author
  • Anthony J. Gill
    Correspondence
    Address for correspondence: Dr Anthony J. Gill, Department of Anatomical Pathology, Royal North Shore Hospital, Pacific Highway, St Leonards NSW 2065, Australia.
    Affiliations
    Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, NSW, Australia

    NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, Sydney, NSW, Australia

    Sydney Medical School, University of Sydney, Sydney, NSW, Australia
    Search for articles by this author
Published:October 11, 2022DOI:https://doi.org/10.1016/j.pathol.2022.07.014

      Summary

      Discontinuous extramural tumour nodules (DTNs) are deposits of metastatic carcinoma in soft tissue not associated with lymph nodes. Although they are established as an adverse prognostic factor in colorectal carcinoma (CRC), under the AJCC eighth edition staging system their presence does not upstage patients who also have lymph node metastases. Counterintuitively, in some situations the presence of lymph node metastases may in effect downstage a patient with DTNs from pN1c to pN1a/b. Therefore, we sought to critically assess the significance of DTNs in a large unselected single institution cohort of patients undergoing surgical resection for CRC.
      Of 3822 CRC patients undergoing surgical resection from 2005 to 2021, DTNs were present in 686 (18.0%). In univariate (HR=2.687, 95CI 2.355–3.065; p<0.001) and multivariate analysis (HR=1.805, 95CI 1.529–2.132; p<0.001) in a model including age, gender, stage, grade, location, lymph node ratio and apical lymph node status, DTNs were associated with worse overall survival (OS). N1c patients (DTN present but no nodal metastasis) demonstrated worse OS compared to the current pN1a group (p<0.001) and were least different to the current pN2a group (p=0.571). Within the current N1a (p=0.013), N1b (p=0.004) and N2a (p=0.002) groups, patients who also had DTNs had worse OS. DTNs were associated with worse OS for all stage III CRCs combined (p<0.001), and for stage IIIB (p<0.001) and stage IIIC (p=0.007) individually.
      We conclude that DTNs are an independent adverse prognostic factor that should be considered in the staging system in a way that is additional to (rather than integrated with) the number of involved lymph nodes. We then assess a simple suggestion for how this could be achieved by increasing the overall stage by one group in the presence of DTNs (requiring the creation of a new stage IIID).

      Key words

      Introduction

      Accurate pathological staging of colorectal carcinomas (CRC) undergoing surgical excision is critical. Staging both guides treatment, most importantly by informing the need for adjuvant chemotherapy, and predicts outcome.
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      The American Joint Committee on Cancer (AJCC) eighth edition 2017 staging system is currently in widespread use.
      It assigns a stage group based on the depth of tumour invasion relative to the layers of the colonic wall (the pT stage), the number of lymph nodes involved by metastatic carcinoma (the pN stage), and the presence or absence of distant metastases (the pM stage).
      Discontinuous extramural tumour nodules (DTNs) can be broadly defined as deposits of carcinoma in soft tissue that are both separate from the main tumour mass and not clearly associated with lymph nodes, but within the lymphatic drainage area of the carcinoma.
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      The precise criteria used to define DTNs in CRC and their impact on staging have changed over time. Under the AJCC fifth edition 1997,
      a soft tissue deposit of CRC greater than 3 mm was classified in the N category (that is as a lymph node metastasis), whereas a deposit up to 3 mm was classified in the T category as tumour extension. Under the AJCC sixth edition 2002,
      a deposit with a smooth contour was classified in the N category, and in the T category (as tumour extension) if the nodule had an irregular contour. Under the AJCC seventh edition 2009,
      DTNs were separately acknowledged under the staging system for the first time. They were considered an independent N group (pN1c) in patients without lymph node metastasis but did not otherwise impact the staging of patients with nodal metastases. Under the current AJCC eight edition 2017,
      the definition of DTNs has been clarified noting that they should also lack histological evidence of residual lymph node or identifiable vascular or neural structures. Therefore, if a vessel wall is identifiable, a tumour deposit should be classified as extramural vascular or lymphatic invasion and if neural elements are identified it should be classified as extramural perineural invasion.
      Despite the slightly different definition, the eighth edition has not changed the impact of DTNs on staging compared to the seventh edition. That is, for patients without lymph node metastases, the presence of DTNs connotes pN1c. In patients with any number of lymph node metastases, the presence of DTNs does not change the N grouping or the overall staging grouping.
      This has led to several situations where the staging appears to be both counterintuitive and not evidence based. For example, if a patient has DTN but no involved lymph nodes they would be considered pN1c. However, if the same patient also had less than four lymph node metastases as well as DTN, they would be effectively ‘downstaged’ to pN1a/b. Also, despite evidence that DTNs are an independent adverse prognostic factor across all stages,
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      the presence of DTNs does not alter the stage group for patients with known lymph node metastases. Finally, there is the issue of patients with low pT stage disease and either a small number of positive nodes or DTN without involved nodes (T1–2N1, T1N2a). Currently these patients are assigned stage IIIA. Stage IIIA CRC patients have been consistently, but perhaps counterintuitively, demonstrated to have a better survival than stage pIIB/C patients.
      • Kim M.J.
      • Jeong S.-Y.
      • Choi S.-J.
      • et al.
      Survival paradox between stage IIB/C (T4N0) and stage IIIA (T1-2N1) colon cancer.
      • Chu Q.D.
      • Zhou M.
      • Medeiros K.
      • Peddi P.
      Positive surgical margins contribute to the survival paradox between patients with stage IIB/C (T4N0) and stage IIIA (T1-2N1, T1N2a) colon cancer.
      • Kim H.S.
      • Kim K.M.
      • Lee S.B.
      • et al.
      Clinicopathological and biomolecular characteristics of stage IIB/IIC and stage IIIA colon cancer: insight into the survival paradox.
      • Li H.
      • Fu G.
      • Wei W.
      • et al.
      Re-evaluation of the survival paradox between stage IIB/IIC and stage IIIA colon cancer.
      It is unclear if this very good survival of patients who are stage IIIA is maintained for patients with DTNs.
      With these issues in mind, we sought to critically assess the significance of DTNs in CRC using our own large and unselected single institution cohort linked to long term survival data. We then sought to evaluate a very simple proposal for how DTNs could be integrated into the staging system for CRC independent of lymph node metastasis. This proposal suggests that the presence of DTNs could simply upstage stage III CRC patients with identical pT and pN status, by one stage group (from stage IIIA to IIIB, IIIB to IIIC, and from IIIC to a new proposed group of IIID).

      Methods

      We searched the computerised database of the Department of Anatomical Pathology Royal North Shore Hospital, Sydney, Australia, for all patients undergoing surgical resection of colorectal carcinoma between 1 January 2005 and 31 December 2021. Exclusion criteria included extracolonic and appendiceal location, tumours undergoing biopsy alone or treated endoluminally only, cases with distant metastatic deposits (stage IV disease), and histological types other than adenocarcinoma as defined by the World Health Organization 2019 system. Rectal tumours were included in the cohort and we did not exclude rectal carcinomas treated with neoadjuvant therapy if they subsequently underwent surgical resection. Cases that had incomplete data for the presence of discontinuous extramural tumour deposits were excluded. When multiple tumours were present, the highest stage tumour was assessed.
      Since 2005 all major resection specimens received in the Department of Anatomical Pathology at the Royal North Shore Hospital have been reported using a structured format recording all pathological data in a synoptic manner.
      • Gill A.J.
      • Johns A.L.
      • Eckstein R.
      • et al.
      Synoptic reporting improves histopathological assessment of pancreatic resection specimens.
      For CRCs this report has always included the presence or absence of DTNs as an explicit datapoint. The precise criteria used to define DTNs were based on whichever AJCC staging system was current at the time as described above. That is, up until 31 December 2009 the sixth edition was used
      ; from 1 January 2010 to 31 December 2017 the seventh edition was used
      ; and from 1 January 2018 onwards the eighth edition was current.
      We emphasise that during this period the impact of DTNs on overall stage changed over the different AJCC editions but the basic definition of a DTN, as a deposit of carcinoma in soft tissue separate from the tumour and not associated with a lymph node, did not. For example, for staging purposes under the sixth edition a DTN altered the N or T category based on whether it had a smooth or irregular contour, whereas under the seventh and eighth edition DTNs only altered the N category (pN1c) if there were no lymph node metastases. However, under both systems the definitions of DTNs were the same. Similarly, under the eighth edition definite association with large vessels, lymphatics or nerves indicated extramural vascular invasion, lymphatic invasion, or perineural invasion, respectively. However, during this period we also recorded all such examples as DTNs in our structured report. For example, using the current approach in our department which is based on the eighth edition, a tumour deposit in soft tissue associated with a vein away from the main tumour mass would be considered both DTN and extramural vascular invasion but would not warrant the pN1c categorisation. Therefore, our definition of DTN did not change significantly over the study period.
      However, the impact of DTNs on the pT, pN and overall stage groups did change over the three different AJCC editions used during the study period, as did the definitions and impact on stage grouping of several other parameters. Therefore, we retrospectively reviewed all cases and assigned an AJCC eighth edition stage for all CRCs. It was the eighth edition staging system that was used for all statistical analysis.
      For tumour location, the right colon was defined as caecum, ascending colon, hepatic flexure, and transverse colon. The left colon was defined as splenic flexure, descending colon, sigmoid, and rectum. The lymph node ratio (LNR) was defined as the total number of involved nodes divided by the total number of nodes resected. Data on adjuvant therapy received (if any) was not available. All cause survival was current as of 1 January 2022. The study was approved by the Northern Sydney Local Health District Human Research Ethics Committee.
      All statistical analyses were performed using SPSS, version 27.
      Overall survival (OS) was defined as time from surgery until any-cause death (date of census 1 January 2022). The continuous LNR data was categorised into two groups: ≤0.4 and >0.4. The effect of DTNs on overall survival was also estimated in subgroup analyses for all stage III cases combined, and for stage III cases by individual subgroup (IIIA, IIIB, IIIC) and for each individual nodal group (pN1a, pN1b, pN2a, pN2b). The Kaplan–Meier method was used to obtain the median overall survival for age, sex, TNM stage, histological grade, tumour location, the presence of DTN, apical node (APN) metastasis status and LNR. Univariate and multivariate cox proportional hazard regression analyses were used to evaluate overall survival in a model that included age, sex, AJCC eighth edition stage, histological grade, tumour location, the presence of DTN, apical lymph node status, and LNR. All statistical tests were two sided, and p values ≤0.05 were considered significant.
      Having explored the impact of DTNs on survival in CRCs staged according to the eighth edition AJCC, we then evaluated one potential proposal for how DTNs could be integrated into a novel staging system for CRC under which they would be considered variables independent of lymph node metastasis. The full proposal is presented in Table 1. Briefly, we suggest that the presence of DTNs could simply upstage stage III CRC patients with identical pT and pN status, by one stage group (from stage IIIA to IIIB, IIIB to IIIC, and from IIIC to a new proposed group of IIID).
      Table 1Comparison of the current AJCC eighth edition staging system
      with our proposal for a modified staging system
      T categoryN categoryM categoryCurrent AJCC eighth edition stage groupProposed stage group if DTN absentProposed stage group if DTN present
      TisN0M000
      T1, T2N0M0II
      T3N0M0IIAIIA
      T4aN0M0IIBIIB
      T4bN0M0IICIIC
      T1–T2N1M0IIIAIIIAIIIB
      T1N2aM0IIIAIIIAIIIB
      T3–T4aN1M0IIIBIIIBIIIC
      T2–T3N2aM0IIIBIIIBIIIC
      T1–T2N2bM0IIIBIIIBIIIC
      T4aN2aM0IIICIIICIIID
      T3–T4aN2bM0IIICIIICIIID
      T4bN1–N2M0IIICIIICIIID
      Any TAny NM1aIVAIVA
      Any TAny NM1bIVBIVB
      Any TAny NM1cIVCIVC
      Under this novel proposal, if the patient has any number of DTNs they are upstaged by one stage group; this would result in a new stage group of IIID.
      AJCC, American Joint Committee on Cancer; DTN, discontinuous extramural tumour nodules.

      Results

      A total of 3822 patients with CRC undergoing surgical resection between 2005 and 2021 and fulfilling the inclusion criteria were identified. The clinicopathological details are summarised in Table 2. Briefly the median age at surgery was 74 (range 21–103 years); 50.9% were female; 759 (19.9%) were stage I, 1522 (39.8%) were stage II, and 1541 (40.3%) were stage III. A total of 686 (17.9%) had DTNs, and DTNs were confirmed absent in 3136 (82.1%). The overall survival of all patients stratified by AJCC eighth edition is presented in Fig. 1. Of note, as consistently reported by others,
      • Kim M.J.
      • Jeong S.-Y.
      • Choi S.-J.
      • et al.
      Survival paradox between stage IIB/C (T4N0) and stage IIIA (T1-2N1) colon cancer.
      • Chu Q.D.
      • Zhou M.
      • Medeiros K.
      • Peddi P.
      Positive surgical margins contribute to the survival paradox between patients with stage IIB/C (T4N0) and stage IIIA (T1-2N1, T1N2a) colon cancer.
      • Kim H.S.
      • Kim K.M.
      • Lee S.B.
      • et al.
      Clinicopathological and biomolecular characteristics of stage IIB/IIC and stage IIIA colon cancer: insight into the survival paradox.
      • Li H.
      • Fu G.
      • Wei W.
      • et al.
      Re-evaluation of the survival paradox between stage IIB/IIC and stage IIIA colon cancer.
      stage IIIA CRC patients demonstrated better long-term survival than stage II patients.
      Table 2Clinicopathological variables and impact on survival in univariate and multivariate analysis for 3822 patients with colorectal carcinoma
      Variablen (%)Median survival (months)
      Median survival calculated using Kaplan–Meier method.
      Univariate analysis
      Cox regression model.
      Multivariate analysis
      Cox regression model.
      HR95% CIpHR95% CIp
      Age (years)
       ≤742021 (52.9)NR11
       >741801 (47.1)86.5212.3402.065–2.651<0.0012.6182.300–2.980<0.001
      Sex
       Female1947 (50.9)156.3751
       Male1875 (49.1)133.5711.0660.946–1.2020.296
      TNM Stage
       I–II2281 (59.7)172.37711
       III1541 (40.3)78.6312.1391.895–2.413<0.0011.3491.151–1.582<0.001
      Histologic grade
       Low2938 (77.4)152.49711
       High856 (22.6)90.5911.7361.520–1.982<0.0011.3131.143–1.509<0.001
      Tumour location
       Right colon1850 (48.7)112.4101.2631.120–1.425<0.0011.1080.951–1.2230.240
       Left colon/rectum1951 (51.3)161.73111
      Discontinuous extramural tumour deposits
       Absent3136 (82.1)164.95111
       Present686 (17.9)46.9392.6872.355–3.065<0.0011.8051.529–2.132<0.001
      Apical lymph node metastasis
       Absent3608 (94.4)152.49711
       Present214 (5.6)28.8833.5162.921–4.231<0.0011.7171.385–2.129<0.001
      Lymph node ratio
       0 to ≤0.43291 (75.8)157.95211
       >0.4531 (48.6)37.6342.5832.129–3.134<0.0011.6641.391–1.989<0.001
      Bold p values are statistically significant.
      CI, confidence interval; HR, hazard ratio; NR, not reached.
      a Median survival calculated using Kaplan–Meier method.
      b Cox regression model.
      Fig. 1
      Fig. 1All cause survival of 3822 colorectal carcinoma patients staged by the AJCC eighth edition staging system.
      Age >74 years, higher stage, high grade, right sided tumour location, the presence of DTNs, apical lymph node involvement and lymph node ratio >0.4 were all significantly associated with worse overall survival in univariate cox-regression analyses (Table 2). In multivariate analysis age >74 years, male gender, stage, grade, DTNs, apical node involvement and lymph node ratio >0.4 maintained statistical significance as adverse prognostic factors for overall survival. However, tumour location did not.
      We then stratified the OS based on the presence or absence of DTNs for the entire cohort (Fig. 2), demonstrating a significantly worse survival in patients with DTNs (p<0.001). Looking specifically at stage III patients (Fig. 3) we demonstrated a significantly worse survival for stage III patients with DTNs as a group (p<0.001). For IIIA patients there was only a trend towards worse survival which did not reach statistical significance (p=0.244, as there were only 164 patients in the group and just 25 with DTNs). For both stage IIIB and stage IIIC patients there was significantly worse survival for patients with DTN (p<0.001 and p=0.007 respectively).
      Fig. 2
      Fig. 2All cause survival of 3822 colorectal carcinoma patients stratified by the presence or absence of discontinuous tumour nodules.
      Fig. 3
      Fig. 3All cause survival for 1541 stage III colorectal carcinoma patients stratified by the presence or absence of discontinuous tumour nodules. (A) All 1541 stage III patients combined (p<0.001); (B) 164 stage IIIA patients (p=0.244); (C) 975 stage. IIIB patients (p<0.001); (D) 402 stage IIIC patients (p=0.007).
      To further examine the appropriateness of the current system under which DTNs do not affect the pN stage unless there are no nodal metastases, we specifically examined the survival effect of DTNs on patients stratified by all nodal stages (Fig. 4). In three of the four nodal groups, the presence of DTNs indicated statistically significantly worse survival: pN1a (p=0.013), pN1b (p=0.004), and pN2a (p=0.002). In the 211 patients with seven or more nodal metastases (pN2b), DTNs just failed to reach statistical significance (p=0.071).
      Fig. 4
      Fig. 4All cause survival of CRC patients with lymph node metastasis by pN stage stratified by the presence or absence of discontinuous tumour nodules. (A) 482 stage pN1a patients (p=0.013); (B) 424 stage pN1b patients (p=0.004); (C) 288 stage pN2a patients (p=0.002); (D) 211 stage pN2b patients (p=0.071).
      We then examined the appropriateness of the change to the eighth edition staging system which resulted in DTNs that clearly represent extramural vascular invasion, lymphatic invasion or perineural invasion no longer resulting in upstaging of patients without lymph node metastases to pN1C. In our cohort, 799 patients were from 2018–2021 and therefore were staged according to the eighth edition. Of these patients, 133 had DTNs present of whom six were stage 2A, eight were stage 3A, 69 were stage 3B, and 50 were stage 3C. That is, the change made to the eighth edition compared to the seventh edition affected only six of 799 patients (0.75%). None of these six patients died during follow-up. However, the mean follow-up of this group was only 7.2 months. Whilst we could potentially argue that the relatively good survival of this cohort (that is behaviour more like stage 3A than 2A) suggests that the seventh edition approach may be more appropriate, there are clearly too few patients and too little follow-up to make any conclusions on this point.
      Finally, we compared the overall survival curves by the current AJCC eighth edition staging system to our novel proposal where we simply suggest that the presence of DTNs should upstage the combined stage by one stage group (Fig. 5). These curves showed very good separation with both 5-year and 10-year median survivals differing by approximately 20% across each of the four proposed stage III subgroups.
      Fig. 5
      Fig. 5Comparison of survival amongst the currently defined AJCC eighth edition stage IIIA, IIIB and IIIC CRCs (A) with our proposal that discontinuous tumour nodules could simply increase the stage by one stage group necessitating the creation of a new stage IIID (B). Under this new proposal the excellent outcomes of stage IIIA is further improved, and all four proposed stages show 5-year and 10-year median survivals that differ by approximately 20%.

      Discussion

      There is level I evidence from multiple studies that DTNs are an independent marker of poor prognosis in CRC.
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      ,
      ,
      • Lord A.C.
      • D'Souza N.
      • Pucher P.H.
      • et al.
      Significance of extranodal tumour deposits in colorectal cancer: a systematic review and meta-analysis.
      Despite this, there remains very significant uncertainty and disagreement about how to best incorporate DTNs into an integrated staging system for CRC. This debate has spanned decades without clear resolution.
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      ,
      • Lord A.C.
      • D'Souza N.
      • Pucher P.H.
      • et al.
      Significance of extranodal tumour deposits in colorectal cancer: a systematic review and meta-analysis.
      • Nagtegaal I.D.
      • Quirke P.
      Tumour deposits in colorectal cancer: the challenge is continuing.
      • Nagtegaal I.D.
      • Tot T.
      • Jayne D.G.
      • et al.
      Lymph nodes, tumor deposits, and TNM: are we getting better?.
      • Nagtegaal I.D.
      • Quirke P.
      Colorectal tumour deposits in the mesorectum and pericolon; a critical review.
      • Ueno H.
      • Mochizuki H.
      • Shirouzu K.
      • et al.
      Multicenter study for optimal categorization of extramural tumor deposits for colorectal cancer staging.
      • Wang S.
      • Guan X.
      • Ma M.
      • et al.
      Reconsidering the prognostic significance of tumour deposit count in the TNM staging system for colorectal cancer.
      • Loughry M.B.
      • Kent O.
      • Moore M.
      • et al.
      Impact on colorectal cancer pathology reporting practice of migration from TNM 5 to TNM 8.
      • Chen V.W.
      • Hsieh M.C.
      • Charlton M.E.
      • et al.
      Analysis of stage and clinical/prognostic factors for colon and rectal carcinoma from SEER registries: AJCC and collaborative stage data collection system.
      • Puppa G.
      • Maisonneuve P.
      • Sonzogni A.
      • et al.
      Pathological assessment of pericolonic tumor deposits in advanced colonic carcinoma: relevance to prognosis and tumor staging.
      It would be naïve to suggest that our single institution study would be able to resolve this controversy and we certainly accept that it is always easier to criticise an existing approach than to develop and validate a better way. However, we hope that we have added useful data to contribute to the debate.
      We confirmed other studies showing that DTNs are an adverse prognostic factor with a significant impact on OS (median survival 165 vs 47 months). This was confirmed in both univariate (HR=2.687) and multivariate (HR=1.805) Cox regression analysis. We then sought to critically address the current AJCC eighth edition approach where the presence of DTNs does not affect pN stage (unless there are no nodal metastasis). Our data demonstrate that in three of the four nodal groups, the presence of DTNs indicated statistically significantly worse survival: pN1a (p=0.013), pN1b (p=0.004), and pN2a (p=0.002). In the 211 patients with seven or more nodal metastasis (pN2b), DTNs just failed to reach statistical significance (p=0.071). Interestingly, the survival curves of patients within a current AJCC eighth edition nodal group who had DTNs were quite similar to those found in the next greater pN stage who did not have nodal metastases (Fig. 4), suggesting that one approach to including DTNs in the staging system could be to simply increase the pN stage as determined by the total number of lymph nodes by one group.
      This finding argues against the current approach and provides strong evidence that the presence of DTNs should be considered in some way additive to the pN stage in patients with nodal disease. Specifically looking at whether it is appropriate for patient with DTNs to be ‘downstaged’ to pN1a/b disease if there are small numbers of involved nodes, we found that AJCC eighth edition pN1c CRCs demonstrated worse OS compared to N1a CRCs (p<0.001) without DTNs. However, their survival was similar to pN1b (p=0.295), and pN2a (p=0.571), being intermediate between the two but more similar to pN2a. N2b demonstrated a significantly worse OS than N1c (p<0.001). That is, we demonstrated that if DTN status is to be integrated into the pN stage, cases with no lymph node metastasis should be categorised somewhere between pN1b and pN2a but are more similar to pN2a.
      We then looked at the specific issue of whether patients with DTNs should remain in the stage IIIA good prognostic group if they have low pT stage disease, and either very few or no positive nodes. Our data were less definitive. Whilst the survival curves separated with an approximately 20% difference at 10 years (Fig. 3), this did not reach statistical significance (p=0.244) because of the low number of patients in the IIIA group (n=164, of whom only 25 had DTNs). The survival curve of these current pIIIA patients with DTNs but no nodal metastases (Fig. 3B) was more similar to the survival curve of the current pIIIB patients without DTNs (Fig. 3C). Although not definitive, we believe these data argue strongly against including patients with DTNs in the very good prognosis pIIIA category. Further, we note that for both stage IIIB and stage IIIC patients, the curves of patients with and without DTNs clearly separated and this was statistically significant (p<0.001 and p=0.007, respectively). From this we can definitively conclude that patients with stage IIIB/C disease and DTNs have a worse outcome than those without DTNs.
      Finally, we explored one potential approach of how DTNs could be integrated into the overall stage groupings. Our proposed approach of increasing the combined group by one category with the creation of a new stage IIID is not perfect. However, the fact that the curves for IIIA, IIIB, IIIC and IIID separated so clearly under this proposal (Fig. 5), with a consistent approximately 20% difference in 5-year and 10-year median survival between each of the groups, certainly indicates that this approach has some merit and warrants further investigation along with other proposals.
      Strengths of this study include the large size (n=3822) of our single institution cohort and linkage to survival data. We have confidence in our survival data and stage determination. It is reassuring that our survival curves when stratified by the standard eighth edition AJCC staging system showed a very similar relationship between stage and OS, as reported by others,
      including the so called ‘stage IIIA paradox’, so that the survival by stage in decreasing order was as follows: IIIA, I, IIA, IIB, IIIB, IIC, IIIC. However, our study was not without its weaknesses. Most importantly we did not have access to treatment data including data on adjuvant therapies received. This is particularly significant because one of the main roles of staging CRC is to guide the need for adjuvant chemotherapy after definitive surgery. Whilst patients with similar stages and similar risk factors would have received similar treatments at any individual time point, these treatments changed over time and we cannot exclude the possibility that as adjuvant therapies improved, the significance of multiple prognostic factors including DTN may have changed. Furthermore, the presence or absence of DTNs would have influenced the risk-benefit decision to offer adjuvant therapy. That is, patients with DTNs but otherwise low risk disease may have been more likely to receive adjuvant therapy than other low risk patients. Secondly, we used overall survival rather than disease free or disease specific survival as our primary endpoint, simply because we did not have access to the latter data. Although overall survival is a robust endpoint, disease free or disease specific survival would have been a better endpoint to more directly assess the impact of DTNs on survival. We did not review the original slides for all 3822 patients in our cohort and rather our ‘real world’ study was based on how DTNs were reported at the time. Whilst it is reassuring that our incidence of DTNs of 18% amongst resected CRC is within the range of 10–22% reported by others,
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      ,
      ,
      • Chen V.W.
      • Hsieh M.C.
      • Charlton M.E.
      • et al.
      Analysis of stage and clinical/prognostic factors for colon and rectal carcinoma from SEER registries: AJCC and collaborative stage data collection system.
      the threshold for reporting DTNs and separating them from lymph node metastases could have changed over time and this may have affected our results. Finally, for the purposes of this project we only considered DTNs as being ‘present’ or ‘absent’ and we did not study the effect of the number of DTNs on prognosis. Current recommendations are that the absolute numbers of DTNs be reported as there is good evidence that the presence of more DTNs correlates with worse prognosis.
      • Nagtegaal I.D.
      • Knijn N.
      • Hugen N.
      • et al.
      Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
      ,
      • Wang S.
      • Guan X.
      • Ma M.
      • et al.
      Reconsidering the prognostic significance of tumour deposit count in the TNM staging system for colorectal cancer.
      In conclusion, our study provides strong evidence that the presence of DTNs is an adverse prognostic factor beyond, and in addition to, the number of involved lymph nodes. We have also illustrated some of the limitations in the way DTNs are currently included in the eighth edition AJCC staging system and provide good evidence that they should be, in some way, considered independent adverse prognostic factors that are additive to the number of involved nodes. Whilst our primary aim was to investigate the issue and provide more data rather than definitively solve the problem, we then provided one example of how DTNs could potentially be integrated into staging by simply increasing the overall stage grouping by one stage. We hope this study informs the discussions during preparation of the upcoming ninth edition staging system.

      Conflicts of interest and sources of funding

      The authors state that there are no conflicts of interest to disclose.

      References

        • Nagtegaal I.D.
        • Knijn N.
        • Hugen N.
        • et al.
        Tumour deposits in colorectal carcinoma: improving the value of modern staging-A systematic review and meta-analysis.
        J Clin Oncol. 2016; 35: 1119-1127
      1. Amin M.B. Edge S.B. Greene F.L. AJCC Cancer Staging Manual. 8th ed. Springer, New York2017
      2. Fleming I.D. Cooper J.S. Henson D.E. AJCC Cancer Staging Manual. 5th ed. Lippincott-Raven, Philadelphia1997
      3. Greene F.L. Page D.L. Fleming I.D. AJCC Cancer Staging Manual. 6th ed. Springer-Verlag, New York2002
      4. Edge S.B. Byrd D.R. Compton C.C. Fritz A.G. Greene F.L. Trotti A. AJCC Cancer Staging Manual. 7th ed. Springer, New York2010
        • Kim M.J.
        • Jeong S.-Y.
        • Choi S.-J.
        • et al.
        Survival paradox between stage IIB/C (T4N0) and stage IIIA (T1-2N1) colon cancer.
        Ann Surg Oncol. 2015; 22: 505-512
        • Chu Q.D.
        • Zhou M.
        • Medeiros K.
        • Peddi P.
        Positive surgical margins contribute to the survival paradox between patients with stage IIB/C (T4N0) and stage IIIA (T1-2N1, T1N2a) colon cancer.
        Surgery. 2016; 160: 1333-1343
        • Kim H.S.
        • Kim K.M.
        • Lee S.B.
        • et al.
        Clinicopathological and biomolecular characteristics of stage IIB/IIC and stage IIIA colon cancer: insight into the survival paradox.
        J Surg Oncol. 2019; 120: 423-430
        • Li H.
        • Fu G.
        • Wei W.
        • et al.
        Re-evaluation of the survival paradox between stage IIB/IIC and stage IIIA colon cancer.
        Front Oncol. 2020; 10595107
        • Gill A.J.
        • Johns A.L.
        • Eckstein R.
        • et al.
        Synoptic reporting improves histopathological assessment of pancreatic resection specimens.
        Pathology. 2009; 41: 161-167
      5. IBM. SPSS Statistics for Macintosh. Armonk: IBM Corp, 2020
        • Lord A.C.
        • D'Souza N.
        • Pucher P.H.
        • et al.
        Significance of extranodal tumour deposits in colorectal cancer: a systematic review and meta-analysis.
        Eur J Cancer. 2017; 82: 92-102
        • Nagtegaal I.D.
        • Quirke P.
        Tumour deposits in colorectal cancer: the challenge is continuing.
        Histopathology. 2008; 52: 768-769
        • Nagtegaal I.D.
        • Tot T.
        • Jayne D.G.
        • et al.
        Lymph nodes, tumor deposits, and TNM: are we getting better?.
        J Clin Oncol. 2011; 29: 2487-2492
        • Nagtegaal I.D.
        • Quirke P.
        Colorectal tumour deposits in the mesorectum and pericolon; a critical review.
        Histopathology. 2007; 51: 141-149
        • Ueno H.
        • Mochizuki H.
        • Shirouzu K.
        • et al.
        Multicenter study for optimal categorization of extramural tumor deposits for colorectal cancer staging.
        Ann Surg. 2012; 255: 739-746
        • Wang S.
        • Guan X.
        • Ma M.
        • et al.
        Reconsidering the prognostic significance of tumour deposit count in the TNM staging system for colorectal cancer.
        Sci Rep. 2020; 10: 89
        • Loughry M.B.
        • Kent O.
        • Moore M.
        • et al.
        Impact on colorectal cancer pathology reporting practice of migration from TNM 5 to TNM 8.
        Histopathology. 2020; 77: 210-222
        • Chen V.W.
        • Hsieh M.C.
        • Charlton M.E.
        • et al.
        Analysis of stage and clinical/prognostic factors for colon and rectal carcinoma from SEER registries: AJCC and collaborative stage data collection system.
        Cancer. 2014; 120: 3793-3806
        • Puppa G.
        • Maisonneuve P.
        • Sonzogni A.
        • et al.
        Pathological assessment of pericolonic tumor deposits in advanced colonic carcinoma: relevance to prognosis and tumor staging.
        Mod Pathol. 2007; 20: 843-855