With interest we read the paper by Loupakis et al. (1), in which the prognostic impact of primary tumor location in metastatic colorectal cancer (CRC) was investigated. They found that patients with left-sided tumors had a lower risk of death and progression as compared with patients with right-sided tumors. We repeated the analyses for metastatic CRC patients and additionally investigated whether survival outcomes are also more favorable for patients with left-sided tumors as compared with right-sided tumors in case of nonmetastatic CRC.
Data were retrieved from the Netherlands Cancer Registry. All CRC patients diagnosed in the Netherlands between 2003 and 2008 with stage I to IV and who underwent resection were included. In line with Loupakis et al. (1), we classified tumors proximal to the splenic flexure as right sided and tumors distal to the splenic flexure as left sided. Crude five-year overall survival (OS) was calculated using Kaplan-Meier curves. OS was defined as the time from diagnosis to death or last follow-up date. Multivariable Cox regression was used to discriminate independent risk factors for death. Additionally, we analyzed recurrence-free survival (RFS) for a subgroup of nonmetastatic CRC patients diagnosed in the southern Netherlands for whom data on diagnosis of distant recurrences were retrospectively collected in 2010 and 2011. RFS was defined as the time from diagnosis to recurrence or death, whichever occurred first, or last follow-up date for patients without recurrence or death.
We identified 7104 patients with metastatic CRC, of whom 4158 (58.5%) had a left-sided tumor. In line with Loupakis et al. (1), we found that patients with left-sided tumors had favorable OS as compared with patients with right-sided tumors (crude 5-year OS 16% vs 10%, adjusted hazard ratio [HR] = 0.80, 95% confidence interval [CI] = 0.76 to 0.84, P < .0001).
We also included 43 482 patients with nonmetastatic CRC, of whom 28 131 (64.7%) had a left-sided tumor. The subgroup consisted of 5457 patients, of whom 3572 (65.5%) had a left-sided tumor. As shown in Table 1, demographic and clinical characteristics differed by tumor location. Table 2 shows the survival outcomes. For the total nonmetastatic CRC population, crude five-year OS was higher for patients with left-sided tumors than for patients with right-sided tumors. However, in multivariable analysis, patients with left-sided tumors had a similar risk of death as compared with patients with right-sided tumors. In the subgroup of patients with data on distant recurrences, crude five-year OS was again higher for patients with left-sided tumors than for patients with right-sided tumors, but crude five-year RFS did not differ between patients with left-sided and right-sided tumors. In multivariable analyses, a trend was observed towards inferior OS for patients with left-sided tumors as compared with patients with right-sided tumors. RFS was inferior for patients with left-sided tumors as compared with patients with right-sided tumors.
Table 1.
Demographic and clinical characteristics of M0 patients by primary tumor location*
| Characteristic | National level |
Regional level with data
on distant recurrences |
||
|---|---|---|---|---|
|
Patients with
right-sided tumors (n = 15 351) No. (%) |
Patients with
left-sided tumors (n = 28 131) No. (%) |
Patients with
right-sided tumors (n = 1885) No. (%) |
Patients with
left-sided tumors (n = 3572) No. (%) |
|
| Sex | ||||
| Male | 6795 (44.3) | 16 230 (57.7) | 841 (44.6) | 2063 (57.8) |
| Female | 8556 (55.7) | 11 901 (42.3) | 1044 (55.4) | 1509 (42.2) |
| P | <.0001 | <.0001 | ||
| Age, y | ||||
| ≤50 | 672 (4.4) | 1899 (6.8) | 87 (4.6) | 233 (6.5) |
| 51–65 | 3218 (21.0) | 9146 (32.5) | 413 (21.9) | 1206 (33.8) |
| >65 | 11 461 (74.6) | 17 086 (60.7) | 1385 (73.5) | 2133 (59.7) |
| P | <.0001 | <.0001 | ||
| Comorbidity | ||||
| None | not available | not available | 513 (27.2) | 1203 (33.7) |
| 1 | 523 (27.8) | 1044 (29.2) | ||
| ≥2 | 733 (38.9) | 1115 (31.2) | ||
| Unknown | 116 (6.1) | 210 (5.9) | ||
| P | <.0001 | |||
| Stage at diagnosis | ||||
| I | 2334 (15.2) | 8164 (29.0) | 275 (14.6) | 1074 (30.1) |
| II | 7471 (48.7) | 10 097 (35.9) | 938 (49.8) | 1313 (36.7) |
| III | 5546 (36.1) | 9870 (35.1) | 672 (35.6) | 1185 (33.2) |
| P | <.0001 | <.0001 | ||
| Mucinous histology | ||||
| Yes | 3356 (21.9) | 2566 (9.1) | 352 (18.7) | 286 (8.0) |
| No | 11 995 (78.1) | 25 565 (90.9) | 1533 (81.3) | 3286 (92.0) |
| P | <.0001 | <.0001 | ||
| Differentiation grade | ||||
| Well/moderate | 10 538 (68.6) | 19 770 (70.3) | 1292 (68.6) | 2534 (70.9) |
| Poor/undifferentiated | 3435 (22.4) | 2774 (9.9) | 451 (23.9) | 336 (9.4) |
| Unknown | 1378 (9.0) | 5587 (19.8) | 142 (7.5) | 702 (19.7) |
| P | <.0001 | <.0001 | ||
| Adjuvant chemotherapy | ||||
| Yes | 3345 (21.8) | 5075 (18.0) | 403 (21.4) | 570 (16.0) |
| No | 12 006 (78.2) | 23 056 (82.0) | 1482 (78.6) | 3002 (84.0) |
| P | <.0001 | <.0001 | ||
| Neo-adjuvant radiotherapy | ||||
| Yes | 10 (0.1) | 8226 (29.2) | 3 (0.2) | 1184 (33.2) |
| No | 15 341 (99.9) | 19 905 (70.8) | 1882 (99.8) | 2388 (66.8) |
| P | <.0001 | <.0001 | ||
| Neo-adjuvant chemoradiation | ||||
| Yes | 6 (0.1) | 1933 (6.9) | 3 (0.2) | 282 (7.9) |
| No | 15 345 (99.9) | 26 198 (93.1) | 1882 (99.8) | 3290 (92.1) |
| P | <.0001 | <.0001 | ||
| Period of diagnosis | ||||
| 2003–2005 | 7255 (47.3) | 13 426 (47.7) | 862 (45.7) | 1738 (48.7) |
| 2006–2008 | 8096 (52.7) | 14 705 (52.3) | 1023 (54.3) | 1834 (51.3) |
| P | .3525 | .0395 | ||
* P values were based on Χ2 tests.
Table 2.
Overall and recurrence-free survival outcomes of M0 patients by primary tumor location
| Demographic level | Overall survival | Recurrence-free survival | ||||
|---|---|---|---|---|---|---|
| Crude 5-y percentage | Hazard ratio (95% CI) | P | Crude 5-y percentage | Hazard ratio (95% CI) | P | |
| National level* | ||||||
| Tumor location | ||||||
| Right-sided | 61 | 1.00 (reference) | .4179 | N/A | N/A | N/A |
| Left-sided | 68 | 0.99 (0.95 to 1.02) | ||||
| Regional level† | ||||||
| Tumor location | ||||||
| Right-sided | 64 | 1.00 (reference) | .0508 | 61 | 1.00 (reference) | .0111 |
| Left-sided | 68 | 1.11 (1.00 to 1.23) | 61 | 1.14 (1.03 to 1.27) | ||
* Hazard ratio adjusted for sex, age, stage at diagnosis, mucinous histology, differentiation grade, adjuvant chemotherapy, neo-adjuvant radiotherapy, neo-adjuvant chemoradiation, and period of diagnosis. P value indicates significance of the hazard ratios. CI = confidence interval; N/A = not available.
† Hazard ratios adjusted for all variables listed under * and additionally for comorbidity.
Contrary to the findings for patients with metastatic CRC, we found on a population-based level that for patients with nonmetastatic CRC left-sided tumor location showed similar OS and unfavorable RFS as compared with right-sided tumor location, even after adjustment for differences in demographic and clinical characteristics between both tumor locations.
Funding
This work was supported by the Netherlands Organisation for Health Research and Development (ZonMw) (grant numbers 113102004, 152002012).
Reference
- 1. Loupakis F, Yang D, Yau L, et al. Primary tumor location as a prognostic factor in metastatic colorectal cancer. J Natl Cancer Inst. 2015;107(3):dju427 doi:10.1093/jnci/dju427. [DOI] [PMC free article] [PubMed] [Google Scholar]