Association of tumour site and sex with survival benefit from adjuvant chemotherapy in colorectal cancer

Summary

Background

Adjuvant chemotherapy can improve 5-year survival in Dukes' C colorectal carcinoma. Improved selection of patients who will respond to adjuvant treatments is required. We investigated whether site of tumour origin, sex, and presence of microsatellite instability (MSI) phenotype were associated with a survival benefit from adjuvant chemotherapy.

Methods

We analysed data for 656 consecutive patients with Dukes' C colorectal carcinoma, with median follow-up of 54 months (range 7–104) and mean age 66·7 years (SD 12·9). We screened tumour samples by PCR for deletions in the BAT-26 mononucleotide repeat to establish MSI status. Details of chemotherapy and survival were obtained by review of hospital and health-department records. Adjuvant chemotherapy (fluorouracil and levamisole) was given with curative intent to 272 (42%) patients.

Findings

Striking survival benefits were seen for patients who had right-sided tumours and who received adjuvant chemotherapy compared with those who did not (48 vs 27% alive at end of study [95% Cl 0·25–0·56], p<·0001), for women (53 vs 33% [0·25–0·56], p<0·0001), and for patients with MSI tumours (90 vs 35% [0·01–0·53], p=0·0007). MSI-positive tumours were slightly more frequent in women than in men (10 vs 7%). Right-sided tumours were more frequently MSI positive than left-sided tumours (20 vs 1%). Men with right-sided tumours benefited from chemotherapy (37 vs 12% [0·24–0·69], p=0·0007) but men with left-sided tumours did not.

Interpretation

The survival benefits seen in patients treated with adjuvant chemotherapy suggest that data from previous trials of adjuvant chemotherapy should be reassessed and the predictive value of MSI status confirmed. Validation of our results will allow better selection of patients for chemotherapy.

Introduction

There are about 10 000 new cases of colorectal carcinoma annually in Australia, of which almost half will eventually be fatal. This disease is the second most frequent cause of cancer death and the second most expensive cancer type to treat.¹ About 40% of cases arise in the right side of the colon—the proximal and transverse segments—with the remainder being located in the distal colon and rectum. Adjuvant chemotherapy can improve 5-year survival of patients with Dukes' C (stage III) colorectal carcinoma,² although the benefits for stage B2 remain controversial.3, 4

10–15% of colorectal carcinomas have widespread somatic alterations in the size of repetitive nucleotide sequences in their DNA.5, 6 This phenotype, referred to as microsatellite instability (MSI), is due to defective DNA mismatch repair, resulting from the inactivation of genes involved in the repair process. Most individuals who have hereditary non-polyposis colon-cancer syndrome have a germ-line mutation in one of these genes, whereas inactivation in most sporadic colorectal carcinomas is thought to result from transcriptional silencing of the hMLH1 repair gene, brought about by DNA methylation.⁷ Colorectal carcinomas with MSI are located mainly in the right-sided colon, frequently have a mucinous and poorly differentiated histology, and frequently show a Crohn's-like lymphoid reaction.8, 9 Many6, 10, 11, 12 but not all13, 14 researchers have reported a better prognosis for tumours with MSI. Because of the short time since the introduction of adjuvant chemotherapy in the management of colorectal carcinoma and the low frequency of MSI, only two reports on the predictive value of MSI have been published. In a mixed tumour series of 508 stage B2 or C colorectal carcinomas, Hailing and colleagues¹⁵ could show no association between MSI and the survival of patients who received chemotherapy. As the investigators stated, however, this finding may have been affected by the study's low power to detect the magnitude of benefit known to be conferred by adjuvant treatment, especially in stage B2 patients. The other study showed prognostic but not predictive value for MSI.¹² Although the number of patients investigated was large, the power to detect predictive value for MSI was decreased because stages of disease were mixed and the chemotherapy status was not known for a third of cases.

We did a retrospective study of the prognostic and predictive value of MSI in patients with Dukes' C colorectal carcinoma treated with or without adjuvant chemotherapy. Several reports have shown differences in the frequencies of MSI5, 6, 10 and various other genetic alterations16, 17, 18, 19 related to tumour site and sex. We tested the hypothesis that these variables could be associated with differences in the response to adjuvant therapy and, therefore, survival benefit from chemotherapy.