Effect of High-Grade Disease on Outcomes of Surgically Treated Colon Cancer

Transcription

1 Ann Surg Oncol DOI /s ORIGINAL ARTICLE COLORECTAL CANCER Effect of High-Grade Disease on Outcomes of Surgically Treated Colon Cancer Ramzi Amri, MD, PhD, Liliana G. Bordeianou, MD, MPH, and David L. Berger, MD Division of General Surgery & Gastrointestinal Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA ABSTRACT Introduction. Tumor grade is one of the cardinal pathological characteristics of colon cancer. Despite a large body of evidence on disease grade in general, the exact impact of high-grade disease in the context of the simplified high/ low-grade dichotomy that is based on glandular formation rate has yet to be quantified. Methods. Patients with sporadic colon cancer treated surgically at our center (24 211) were included in an institutional review board-approved database. We measured the rates of distant and nodal disease spread in baseline pathology and the multivariable hazard radio (mhr) of recurrence and overall- and disease-specific mortality. Results. Among 922 patients with specified tumor grade in baseline surgical pathology, 175 (19. %) had high-grade disease. These patients were at far higher risk of lymph node metastasis (63.8 vs %; P \.1) and metastatic presentation (31.4 vs %; P \.1). These baseline differences also led to significantly worse outcomes, including disease recurrence (17.1 vs. 1.6 %; mhr = 1.83; P =.26), overall mortality (57.7 vs %; mhr = 1.65; P \.1), and colon cancer-specific mortality (39.4 vs %; mhr = 1.57; P =.4). Most significantly, in stage II patients (n = 294), those with high-grade disease (16. %) had an mhr of 2.84 (P \.1) for mortality. Conclusions. High-grade disease on baseline surgical pathology is associated with a considerably higher rate of nodal and distant metastasis in colon cancer. As a result, Presented in abstract form at the 1th annual Academic Surgical Congress, Las Vegas, Nevada, February 3 5, 215. Society of Surgical Oncology 215 First Received: 23 April 215 D. L. Berger, MD dberger@mgh.harvard.edu the colon cancer-related mortality doubles for patients with high-grade disease. These findings were independent of baseline staging and confirm that the high-/low-grade tumor dichotomy is an important prognostic factor greatly influencing colon cancer outcomes across stages. Histologic grade is one of the cardinal tumor characteristics described in the surgical pathological assessment of any resected malignancy. Primary disease classified as high grade is considered to intrinsically have a negative impact on the eventual outcomes of the disease. In colon cancer, disease grade has been part of the standard recommended pathological assessment for decades, and in the past decade, it has transitioned to standardized evaluation criteria. 1 3 The body of evidence from the past century led the College of American Pathologists to recommend that disease-grade be simplified into a two-tiered low-/high-grade system in their 2 consensus statement. 4 The expectation is that disease grade will correlate with prognosis and outcomes. Therefore, disease grade has been subject to several attempts to incorporate it into multifactorial models meant to further individualize risk stratification for patients with colorectal cancer. 5 7 However, in the interim, specific evidence demonstrating the exact impact of the high/low-grade disease dichotomy on both general tumor characteristics as well as the resulting colon cancer outcomes remains sparse. Therefore, this study was designed to measure the magnitude of the influence of high-grade disease as well as its stage-independent effect in colon cancer. METHODS Patients All colon cancer patients treated surgically at our center from 24 to 211 (n = 171) were extracted from our

2 R. Amri et al. institutional cancer registry to be included in a prospectively maintained survival and outcomes database after institutional review board approval. Data on patients were collected from the research patient data repository, the social security death index, as well as patient records from our healthcare network. Due to the significant differences in treatment approach and tumor biology, we exclusively focused on colon cancer and did not include patients with tumors of the rectum. Tumors of the colon were defined as any tumor proximal to the rectosigmoid junction. 8 Additionally, patients with preexisting high-risk disease forms, including inflammatory bowel disease (n = 38), hereditary nonpolyposis colon cancer (n = 19), and familiar polyposis (n = 3) were excluded from analysis. From the remaining population, identifiable disease grade in the surgical specimen (n = 922) were included for final analysis as described in this paper. Excluded were 9 patients without disease grade description in the surgical pathology report. These were either patients with intramucosal (i.e., carcinoma in situ) disease (n = 23), patients without residual disease in the colonic resection specimen (n = 58), or patients in whom disease grade was irrelevant or impossible to determine (n = 9). Surgical Pathology: Definition of High Grade A group of dedicated gastrointestinal pathologists determined the degree of differentiation for all resection specimens of (suspected) adenocarcinomas of the colon, basing their assessment on the degree of gland formation. This is in accordance to the current recommendations by the College of American Pathologists and World Health Organization, which is to use the percentage of the tumor that demonstrates formation of glands or gland-like structures as the only feature by which grade is assessed. 4,9 When a two-tiered grading system is used, the schism between low and high grade lies at a 5 % of the tumor showing gland formation. Low-grade tumors have C5 % glandular formation (well or moderately differentiated adenocarcinomas) and high-grade tumors have \5 % glandular formation (poorly differentiated or undifferentiated adenocarcinomas). By definition, mucinous and signet-cell carcinomas are always considered high-grade. 9 Statistical Analysis All statistical analysis was performed using SPSS (IBM SPSS Statistics for Windows, Version 22.. Release 213. Armonk, NY: IBM Corp.). A two-tailed P value\.5 was considered as the threshold for statistical significance. Statistical significance of distribution differences in dichotomous variables was assessed using a Chi square test (v 2 ), whereas for ordinal values this was done using a Mann Whitney U test. We measured the relative risk (RR) of patients with high-grade disease having concurring negative surgical and pathological outcomes, which also allows giving a 95 % confidence interval for these characteristics. Outcomes analyzed were recurrence, overall metastasis, and overall and disease-specific mortality, expressed as percentage outcomes and hazard ratios (HR) calculated through timestandardized Cox proportional hazards regressions. Additionally, in multivariable analysis, Cox regression was used to assess the aforementioned effects in the form of multivariable hazard ratios (mhr) adjusted for tumor-, node-, metastasisstage or AJCC staging, age and Charlson comorbidity score not including colon cancer, where relevant. Selection of any other potentially relevant covariables was performed using backward elimination accounting for model fit (R 2 ) and fit of the candidate covariables in the model. RESULTS Of a total of 922 patients with specified tumor grade included for analysis, 175 (19. %) patients had high-grade disease on baseline pathology, with the 747 remaining patients composed of 735 (79.7 %) low-grade cases and a subset of 12 patients with tumors (1.3 %) reported as intermediate-grade. Baseline Characteristics General demographics (Table 1) of patients with highgrade disease, including age, gender, and ethnicity, were not significantly different from the remainder of the cohort. Overall comorbidity rates were largely similar, except for a higher mean ASA score (2.48 vs. 2.35; P =.15). Highgrade patients did have a significantly lower rate of screening diagnosis (9.7 vs %; P \.1) despite similar rates of a history of colorectal polyps and earlier diagnosis of colorectal cancer. High-grade patients also were twice as likely to present with metastatic disease, as their rate of confirmed metastatic disease within 3 days of the index admission amounted to 31.4 versus 15.8 % in other patients (P \.1). Perioperative Characteristics Perioperative characteristics are displayed in Table 2. Patients with high-grade disease were more likely to have undergone a right-sided colectomy (66.9 vs. 5.2 %; P \.1) and less likely to undergo a (recto-)sigmoidectomy (13.1 vs. 28. %; P \.1). During surgery, these patients also were more likely to have clinically T4 tumors (2.6 vs %; P =.3) and had markedly

3 High-Grade Colon Cancer TABLE 1 Baseline characteristics High-grade disease P value Age (median, IQR) 7 (2) 67 (21).71 Gender (% male) 45.1 % 52.2 %.92 Ethnic minority (% nonwhite) 11.4 % 1.2 %.63 Smoker (% active) 1.9 % 11.4 %.84 Charlson comorbidity (mean ± SD) a.78 ± ± ASA score (mean ± SD) 2.48 ± ± History of colorectal polyps (%) 12. % 12.2 %.95 Previous CRC (%) 1.7 % 2.4 %.58 Screening diagnosis (%) b 9.7 % 25.8 % \.1 Baseline metastasis (%) c 31.4 % 15.8 % \.1 Neoadjuvant chemotherapy (%) 3.4 % 3.7 %.84 Palliative case (%) 3.4 % 1.9 %.24 IQR interquartile range, SD standard deviation, CRC colorectal carcinoma a Charlson comorbidity score without accounting for colon cancer b Includes screening colonoscopy, fecal occult blood testing, and follow-up for preexisting polyposis c Established preoperatively through imaging and confirmed within 3 days of index surgery TABLE 2 Associations of high-grade disease with perioperative characteristics High-grade disease RR (95 % CI) P value Procedure duration (min, median, IQR) 129 (12) 123 (1).76 Laparoscopic procedure (%) 2.6 % 26.5 %.77 ( ).1 Type of procedure Left-sided 1.9 % 12.7 %.85 ( ).5 Right-sided 66.9 % 5.2 % 1.33 ( ) \.1 Sigmoid 13.1 % 28. %.47 (.32.7) \.1 Other (total) 9.1 % 9.1 % 1. ( ).99 Multivisceral resection (%) 16. % 14.5 % 1.11 ( ).61 Intraoperative observations Adhesive (ct4b) tumor 2.6 % 11.9 % 1.73 ( ).3 Lymphadenopathy 8. % 2.7 % 2.99 ( ).1 Metastasis 16. % 7.4 % 2.17 ( ) \.1 Admission (days, median, IQR) 6 (5) 5 (4).4 In-hospital mortality (%) 5.1 %.8 % 6.4 ( ) \.1 3-day mortality 5.7 % 1.4 % 3.88 ( ).2 Postoperative chemotherapy 48.6 % 36.4 % 1.33 ( ).3 Lymphadenopathy and metastasis visualized: intraoperatively established through visual and palpatory assessment reported in surgical note higher perioperative visual identification rates of lymphadenopathy (8. vs. 2.7 %: RR 2.99; P =.1) and metastasis (16. vs. 7.4 %, RR 2.17; P \.1). The ensuing admission for patients with high-grade disease had a median admission duration that was longer by 1 day (6 vs. 5 days; P =.4). The mortality during the index admission also was considerably higher at 5.1 versus.8 %; five of nine cases were related to respiratory issues

4 R. Amri et al. TABLE 3 Association of high-grade disease with baseline pathology High-grade disease RR (95 % CI) P value T stage (%) T1 3.5 % 15.9 %.22 (.9.48).2 T2 6.3 % 14.1 %.45 (.25.81).9 T % 48.7 %.99 ( ).97 T % 21.3 % 1.96 ( ) \.1 N stage (%) N 37.7 % 6.4 %.63 (.51.76) \.1 N % 26.2 %.98 ( ).89 N % 13.4 % 2.73 ( ) \.1 AJCC stage III (% N1 or N2) 63.8 % 39.6 % 1.61 ( ) \.1 AJCC stage IV (% M1) 18.3 % 8.2 % 2.24 ( ) \.1 Small vessel disease (%) 7.8 % 35.3 % 2. ( ) \.1 Large vessel disease (%) 52.4 % 28.7 % 1.83 ( ) \.1 Perineural invasion (%) 41.8 % 23.1 % 1.8 ( ) \.1 Infiltrating tumor border configuration (%) 71.3 % 66.8 % 1.7 ( ).24 peritumoral lymphocytic response (%) 63.6 % 77.3 %.82 (.73.93).2 Extramural vascular invasion (%) 57.1 % 27.8 % 2.5 ( ) \.1 Radial margin involvement (%) 12.6 % 4.6 % 2.76 ( ) \.1 (pneumonia/ards), and the remaining four respectively due to stroke, liver failure, failure to thrive, and terminal disease-related withdrawal of care. The admission duration and morality rates appear to confirm that at least a portion of these high-grade patients had a considerably worse performance status, which was objectified in baseline characteristics by the higher mean ASA scores. Surgical Pathology Characteristics in postop surgical pathology are shown in Table 3. Patients with high-grade disease had significantly more advanced disease and more aggressive characteristics (all P \.1), including T4 disease (41.7 vs %), lymph node metastasis (63.8 vs %), M1 disease (18.3 vs. 8.2 %), small vessel disease (7.8 vs %), large vessel disease (52.4 vs %), perineural invasion (41.8 vs %), and extramural vascular invasion (57.1 vs %). Additionally, patients with high-grade disease had a threefold increase of radial margin involvement (12.6 vs. 4.6 %; P \.1). Outcomes The differences in baseline staging also led to a significantly higher risk of poor outcomes (Table 4). These included metastatic recurrence (overall: 17.1 vs. 1.6 %, stage-adjusted mhr: 1.83; P =.6; stage III: 28. vs %; mhr = 1.71; P =.42), overall metastasis (48.6 vs %, P \.1, stage-adjusted mhr: 1.28; P =.61), overall mortality (57.7 vs %; age-, comorbidity- and stage-adjusted mhr: 1.65; P \.1), and colon cancer-specific mortality (39.4 vs %; stageand comorbidity-adjusted mhr: 1.57; P =.4). Utility for Risk Stratification in Stage II and Stage III Patients Of the included patients, nearly a third (n = 294; 31.9 %) were AJCC stage II. Among those, high-grade disease was present in 47 patients (16. %). Figure 1 illustrates the strong links that this had with survival. In 5- year survival estimates, the difference in survival is considerable, at a stage-appropriate 76.5 % for patients with low-grade disease compared with 53.2 % for those with high-grade disease (P \.1). 1 A multivariable Cox proportional hazards model shows that this effect is largely independent of other significant predictors of mortality, including patient age, T-stage, and comorbidity expressed through a Charlson comorbidity score (not including colon cancer in the score).

5 High-Grade Colon Cancer TABLE 4 Outcomes, including stage-stratification and multivariable analysis High-grade P value Multivariable Cox regression P value Covariable(s) HR (95 % CI) Follow-up (months, median, IQR) 25 (43) 45 (47) \.1 Disease-free survival (months, median, IQR) 7 (42) 36 (58) \.1 Survival (months, median, IQR) 29 (44) 47 (46) \.1 Metastatic recurrence 17.1 % 1.6 %.15 AJCC 1.83 ( ).6 Stage I % (/13) 3.2 % (6/188).97 Stage II 14.9 % (7/47) 12.1 % (3/247).6 T-st 1.65 ( ).23 Stage III 28. % (23/82) 17.6 % (43/244).42 T-st, N-st 1.71 ( ).42 Metastasis, overall 48.6 % 26.4 % \.1 AJCC 1.28 ( ).61 Stage I % (/13) 4.3 % (9/188).44 Stage II 19.1 % (9/47) 14.2 % (35/247).38 T-st 1.7 ( ).16 Stage III 53.7 % (44/82) 38.1 % (93/244).14 T-st, N-st 1.27 ( ).21 Mortality, overall 57.7 % 33.3 % \.1 AJCC, age, ch 1.65 ( ) \.1 Stage I 7.7 % (1/13) 2.2 % (38/188).51 T-st, age, ch.23 ( ).15 Stage II 46.8 % (22/47) 23.5 % (58/247).1 T-st, age, ch 2.84 ( ) \.1 Stage III 58.5 % (48/82) 43. % (15/244).15 T-st, N-st, age, ch 1.36 ( ).83 Stage IV 93.8 % (3/32) 78.7 (48/61).61 T-st, N-st, age, ch 1.66 ( ).43 Mortality, colon cancer 39.4 % 16.9 % \.1 AJCC, Ch 1.57 ( ).4 Stage I % (/13) 6.4 % (12/188).35 Stage II 14.9 % (7/47) 7.7 % (19/247).11 T-stage 2.31 ( ).11 Stage III 45.1 % (38/82) 29.9 % (73/244).12 T-st, N-st 1.52 ( ).51 Stage IV 78.1 % (25/32) 72.1 % (44/61).53 T-st, N-st 1.19 (.7 2.5).52 Covariables AJCC: 7th edition American Joint Committee on Cancer classification Ch Charlson comorbidity score disregarding colon cancer, T-st T-stage, N-st N-stage DISCUSSION Context of the Study and Findings A large body of evidence from the 198s and 199s featured the prognostic value of disease grade, which was usually stratified into four levels at that time (well, moderately, poorly differentiated, and undifferentiated). Most research confirmed that advanced disease grade in colon cancer pathology correlated with negative outcomes, most often independent of staging However, the thencommon four-level assessment was subject to high levels of interobserver variation and was hence considered too subjective. 2,14 Therefore, in their late 1999 consensus statement, the College of American Pathologists recommended a move toward a two-tiered grading system collapsing disease grade into low and high grade, recommending the use of a 5 % rate of gland formation of the specimen as a threshold between these grades. 4 This recommendation was implemented and became part of standard protocols in the following years. 3 Perhaps due to the intuitive expectation that high-grade disease leads to worse outcomes, exact estimates for the specific impact of high-grade disease in the context of this dichotomized grading system are scarce. By reviewing our cohort of colon cancer patients, several high-risk characteristics and complicating factors in surgery were shown to be associated with high-grade disease. High-grade disease was disproportionately more prevalent in right-sided tumors and was associated with a 6 % higher chance of encountering an adhesive tumor in addition to higher intraoperative visualization rates of lymphadenopathy and metastasis, as well as higher perioperative mortality. Baseline characteristics, which show a similar rate of preexisting polyposis in high-grade patients combined with considerably lower rates of screening diagnosis, seem to show that these patients developed more interval carcinomas, with a right-sided predominance in surgical pathology, possibly suggesting that a disease mechanism different from the classic adenoma-adenocarcinoma or the serrated pathway may be at play in these tumors. 15,16

6 R. Amri et al. Kaplan-Meier survival curve: Overall survival (n=294) 76.5% vs. 53.2%; P=.1 3-year survival (n=191) 87.6% vs. 63.5%: P<.1 5-year survival (n=111) 81.5% vs. 38.1%: P<.1 Cumulative Survival (%) P<.1 Numbers at risk High-grade n high-grade Survival (years) Multivariable Cox regression: Outcome: mortality Reference group: high grade disease Hazard Radio (HR)=1 High grade disease: HR: % Cl: Covariables: Age in years (HR: 1.6; P<.1) T-stage (3,4a, 4b) (HR: 1.57; P=.2) Charlson score (HR: 1.32; P<.1) Survival (%) P< Survival (years) FIG. 1 Impact of high-grade disease in Stage II patients on survival. Yellow lines depict high-grade disease, blue lines non-high grade High-grade disease also was shown to be associated with a considerably higher stage at presentation with over twice the rate of baseline metastasis. As a result, the colon cancer-related mortality doubles for patients with highgrade disease. More interestingly, this effect was shown to remain strong after adjustment for baseline staging, reinforcing suggestions that high tumor grade is an important stage-independent predictive factor for survival. We also demonstrate its utility as a tool for risk stratification within AJCC stages, the most relevant of which is stage II disease: In these patients, the identification of high-risk features is of particular importance in trying to decide for or against adjuvant chemotherapy. To complicate matters even more, several population studies have shown that certain subgroups of stage II patients can have similar or even worse survival than stage III subsets (e.g., stages IIB/IIC vs. IIIA and IIB vs. IIIA). 17,18 These premises underline the importance of identifying factors that will be able to discern the higher-risk subset of stage II patients who would benefit from adjuvant treatment.

7 High-Grade Colon Cancer Limitations and Further Research The findings in this paper are based on postoperative pathology reports and therefore were not in any way subject to blinding or validation by reevaluation or interrater agreement assessment. Interrater variability is a known issue in the assessment of disease grade. 2,19 However, the use of the high/low dichotomy cancels out one of the largest sourced of inter-rater variation by clustering G1 (welldifferentiated) and G2 (moderately differentiated) tumors together into a single low-grade category. 2 Another important mitigator of the risk of bias is that the discussed findings are not influenced by or known to the pathologist at the moment of assessment, nor is there any intrinsic reason for pathologists to be biased. Microsatellite instability and mutation analysis were performed sporadically in this cohort and in numbers too small to yield any significant correlations. Further research should explore relationships between high-grade disease forms with genetic mutation profiles, especially with the goal to uncover specific related pathways. CONCLUSIONS The prime clinical implications of the demonstrated risks associated with high-grade disease are its stage-independent effect on outcomes. High-grade disease also was shown to be independently associated with more difficult surgery and worse postoperative outcomes. If detected in preoperative assessment, high-grade disease needs to be weighed in the decision to operate in advanced cases with low chances of cure. With regards to its impact on longterm outcomes, the hazard ratios provided also should play an important role in estimating the chances of recurrence and in any decisions regarding adjuvant treatment for Stage II patients with high-grade disease. DISCLOSURE REFERENCES The authors report no conflict of interest. 1. Chapuis PH, Dent OF, Fisher R, et al. A multivariate analysis of clinical and pathological variables in prognosis after resection of large bowel cancer. Br J Surg. 1985;72(9): Blenkinsopp WK, Stewart-Brown S, Blesovsky L, Kearney G, Fielding LP. Histopathology reporting in large bowel cancer. J Clin Pathol. 1981;34(5): Washington MK, Berlin J, Branton P, et al. Protocol for the examination of specimens from patients with primary carcinoma of the colon and rectum. Arch Pathol Lab Med. 29;133(1): Compton CC, Fielding LP, Burgart LJ, et al. Prognostic factors in colorectal cancer. College of American Pathologists Consensus Statement Arch Pathol Lab Med. 2;124(7): Weiser MR, Gonen M, Chou JF, Kattan MW, Schrag D. Predicting survival after curative colectomy for cancer: individualizing colon cancer staging. J Clin Oncol. 211;29(36): Manilich EA, Kiran RP, Radivoyevitch T, Lavery I, Fazio VW, Remzi FH. A novel data-driven prognostic model for staging of colorectal cancer. J Am Coll Surg. 211;213(5): Greene FL, Stewart AK, rton HJ. A new TNM staging strategy for node-positive (stage III) colon cancer: an analysis of 5,42 patients. Ann Surg. 22;236(4): discussion Li M, Li JY, Zhao AL, Gu J. Colorectal cancer or colon and rectal cancer? Clinicopathological comparison between colonic and rectal carcinomas. Oncology. 27;73(1 2): Hamilton SR, Aaltonen LA. Pathology and genetics of tumours of the digestive system. Lyon: International Agency for Research on Cancer; Siegel R, Desantis C, Jemal A. Colorectal cancer statistics, 214. CA Cancer J Clin. 214;64(2): Newland RC, Dent OF, Lyttle MN, Chapuis PH, Bokey EL. Pathologic determinants of survival associated with colorectal cancer with lymph node metastases. A multivariate analysis of 579 patients. Cancer. 1994;73(8): Wiggers T, Arends JW, Volovics A. Regression analysis of prognostic factors in colorectal cancer after curative resections. Dis Colon Rectum. 1988;31(1): Deans GT, Patterson CC, Parks TG, et al. Colorectal carcinoma: importance of clinical and pathological factors in survival. Ann R Coll Surg Engl. 1994;76(1): Thomas GD, Dixon MF, Smeeton NC, Williams NS. Observer variation in the histological grading of rectal carcinoma. J Clin Pathol. 1983;36(4): Hamilton SR. The adenoma-adenocarcinoma sequence in the large bowel: variations on a theme. J Cell Biochem Suppl. 1992;16G: Patai AV, Molnár B, Tulassay Z, Sipos F. Serrated pathway: alternative route to colorectal cancer. World J Gastroenterol. 213;19(5): Gunderson LL, Jessup JM, Sargent DJ, Greene FL, Stewart AK. Revised TN categorization for colon cancer based on national survival outcomes data. J Clin Oncol. 21;28(2): O Connell JB, Maggard MA, Ko CY. Colon cancer survival rates with the new American Joint Committee on Cancer sixth edition staging. J Natl Cancer Inst. 24;96(19): Jass JR, Atkin WS, Cuzick J, Bussey HJR, Morson BC, rthover JMA, Todd IP. The grading of rectal cancer: historical perspectives and a multivariate analysis of 447 cases. Histopathology. 1986;1(5): Compton CC. Optimal pathologic staging: defining stage II disease. Clin Cancer Res. 27;13(22 Pt 2):6862s 7s.