Prediction and assessment of pathological complete response following neoadjuvant chemoradiotherapy for locally advanced rectal cancer

© 2018 Jennifer Ellen RyanIntroduction The management of patients with rectal cancer who develop a pathological complete response (pCR) following neoadjuvant chemoradiotherapy (nCRT) presents an ongoing challenge to clinicians. Some authors have suggested that the presence of a clinical complete response may allow patients to be spared the morbidity and potential risks of surgery through adoption of a ‘watch and wait’ policy with surgery only in the setting of clinical failure. However clinical response and pathological response are not always wellcorrelated and widespread adoption of this regimen is limited by accurate methods to assess and confirm the presence of a pCR without a surgical specimen. Method A systematic review of the literature has been performed to identify methods to predict and assess a pathological complete response to nCRT. The first part of the literature review focuses on factors predictive of a pCR, specifically clinical features, radiological features and histological and molecular markers. The second part of the literature review aims to determine methods to accurately assess the presence of a pCR following neoadjuvant treatment. Following this an institutional database was interrogated to determine clinical and radiological features associated with prediction and assessment of a pCR and a multimodal predictive model has been developed. Molecular analysis has been performed to identify genetic influences on pCR. The role of radiological imaging in the assessment of pCR will be explored and the prognostic and clinical significance of metabolic response assessment by 18F FDG PET CT has been investigated in detail. Results Assessment Histology and clinical assessment remain the most effective methods of assessment of pCR with histology considered the gold standard. Clinical assessment is limited to low rectal tumours and is open to significant inter-rater variability while histological examination requires a surgical specimen for accurate assessment. Radiological assessment of pCR demonstrates the greatest potential for assessment of pCR without the need for surgery with diffusion weighted MRI and 18F FDG PET CT providing the greatest accuracy. It is likely that improved accuracy will be achieved with multimodal assessment of response combining the benefits of clinical, serological, endoscopic and radiological methods of response assessment. Prediction Clear methods to predict pCR prior to the commencement of therapy have not been defined. Clinical and radiological features of the primary cancer have limited ability to predict response. Molecular signatures hold the greatest potential to predict response however adoption of this technology is limited by poor concordance of biomarkers between cohorts. Conclusion Accurate prediction and determination of a complete pathological response is paramount if a non-operative approach is to be undertaken with confidence in oncological outcomes. A variety of methods are available but currently they lack sufficient sensitivity and specificity to define management. Despite a large volume of research the ability to predict which patients are likely to sustain a pCR and accurately assess those patient who have a pCR remains elusive. Further research into models incorporating both prediction and assessment into decision-making is required