Emerging trends in the prediction of pathological tumour response in rectal cancer following neoadjuvant therapy.

Abstract

The treatment landscape of locally advanced rectal cancer (LARC) is moving towards total neoadjuvant therapy and potential organ preservation. Multimodality neoadjuvant therapy (NAT) and surgery remain the standard-of-care for LARC, with consideration for adjuvant chemotherapy. Neoadjuvant chemotherapy alone is another emerging strategy.1 Due to variation in multimodality treatment choice, biomarkers to predict response to NAT could personalize treatment plans for patients.2 Of particular interest are predictors of pathologic complete response (pCR) to guide selection for a watch and wait approach, or identify those who would benefit from surgery due to poor response to NAT or a higher risk of disease relapse. Tumour response to NAT is a complex interplay between variables including tumour biology, patient factors, and the sequencing and timing of the selected NAT approach. Presently available modalities (MRI and endoscopy) can aid assessment but do not predict for tumour response. There are currently no clinically actionable biomarkers to predict NAT response, however several promising contenders are emerging.2 Circulating tumour DNA (ctDNA) has been shown as a risk-stratification tool to guide adjuvant therapy choice in patients following colon cancer resection.3 The presence of ctDNA following long course chemoradiation and surgery is an adverse prognostic factor.4 Evidence for ctDNA in the neoadjuvant setting for LARC is also evolving. As a marker for minimal residual disease, studies have demonstrated that detectable ctDNA predicts for poorer response to NAT and potentially identifies patients that should proceed to surgery rather than a watch and wait strategy.5, 6 Other biochemical markers such as microRNA and tumour-infiltrating lymphocytes are also being evaluated.7, 8 The gut microbiome has a role in the carcinogenesis of colorectal cancer, and potentially predicts response to anticancer therapy through the effect of certain bacteria on pharmacokinetics, pharmacodynamics and immune response.9 However, questions remain regarding the specific microbiome biomarkers that can predict LARC response to NAT and its utility in clinical practice is currently limited.9 Immunogenomics is becoming increasingly relevant to the management of LARC, with immune checkpoint inhibitors exhibiting high response rates in mismatch repair deficient (dMMR) tumours and 100% clinical complete response in dMMR LARC.10 Initiatives such as the Cancer Genome Atlas Program are genomically sequencing colorectal cancers to identify genomic alterations beyond microsatellite instability that may predict responsiveness to immunotherapy and the likelihood of a pCR.11 Radiomics is a process whereby disease features, or ‘radiological biomarkers’ are extracted from standard patient imaging, constructing an algorithm which can then aid in predicting treatment response.12 T2 weighted MRI-based software for LARC successfully predicts pCR and nodal status post-NAT.13 Combined radiomics with both multi-parameter MRI (T1/T2/Contrast enhanced) and functional PET-CT may provide better predictive value for pCR.14 Currently, the widespread adoption of radiomics has been hindered by the technological and cost burden, limited reproducibility of data, and time-intensive manual data extraction.13 Larger prospective trials are required to validate its use in LARC management.13, 15 With the adoption of Artificial Intelligence (AI) and deep learning, data extraction will become increasingly accessible and construct reproducible algorithms,13 making clinical application feasible. Body composition measurement based on standard CT-staging is another ‘radiological biomarker’.16 Body composition is more accurate than body surface area for systemic therapy dosing, and predicting poor tolerance to chemotherapy in colorectal cancer.17 Recent evidence has suggested that skeletal muscle area pre-NAT can predict tumour response and survival outcomes for LARC, presenting a promising area for future studies.16 In summary, multiple promising biomarkers are emerging to identify the optimal individualized treatment paradigm for patients with LARC. It is hoped that further research and increasing validation of new technologies such as AI, will tailor future personalized decision making for patients with LARC. Open access publishing facilitated by The University of Melbourne, as part of the Wiley - The University of Melbourne agreement via the Council of Australian University Librarians. Matthew Y. Wei: Writing – original draft; writing – review and editing. Yasser Arafat: Validation; writing – review and editing. Margaret Lee: Validation; writing – review and editing. Suzanne Kosmider: Validation; writing – review and editing. Matthew Loft: Validation; writing – review and editing. Ian Faragher: Validation; writing – review and editing. Peter Gibbs: Conceptualization; validation; writing – review and editing. Justin Yeung: Conceptualization; supervision; writing – review and editing.