Tuberculosis and cancer: a complex and dangerous liaison

520 www.thelancet.com/oncology Vol 12 June 2011 guidelines suggests that from the patients’ perspective, these sequelae are equally important. This study has other limitations. Although the secondary analyses of CRM-negative patients were planned in the trial design, multiple analyses of the same data, even if prespecifi ed, create a risk for false positive fi ndings—particularly in a subset of 435 (23%) of 1861 patients of the total population. Moreover, stage III cancers are more advanced and are likely to cause more diffi cult surgery. The authors do not account for the quality of surgery nor the quality of the surgical specimen based on the completeness of the removal of the mesorectum, as described in the CRO7 study. How do the results fi t with present knowledge? Results from the Swedish rectal cancer trial originally also showed a 10% absolute improvement in survival. Results from the CR07 trial suggested SCPRT was of benefi t in terms of lowering local recurrence for all tumour locations, all pathological stages, and good, average, or poor quality surgery. The Dutch study predates the routine use of MRI, which can predict extramural spread and a free CRM (>1 mm) in 95% of patients, yet we still have no accurate method of clinically staging lymph nodes. How can this study guide treatment in future? TME reduced the rate of local recurrence (even without radiotherapy) to below 10%, but highlighted a 30–40% risk of metastatic disease. The choice of treatment, which ranges from no treatment to treatment intensifi cation with chemotherapy before chemoradiation, will depend on patient-related factors, tumour location, and results of high spatial resolution MRI imaging. Encouraging as the subgroup analysis is for stage III patients, the benefi ts are not clear-cut for all patients, and point to a need to develop strategies to select cases better for the use of SCPRT and chemoradiation.