Predicting RadIotherapy ReSponse of Rectal Cancer With MRI and PET

Bowel cancer is the second most prevalent cancer in Australia with rectal cancer the most common subgroup, representing 5% of all cancer diagnoses in 2008 [1]. Rectal cancer also accounts for 4.6% of all cancer deaths in New South Wales [1]. All rectal cancer that is T3/4 or node positive on pre-operative assessment should be considered for pre-operative therapy, which has been shown to significantly improve local control compared to postoperative chemoradiation. [2]. Pre-operative chemoradiation is standard pre-operative treatment at the Northern Sydney Cancer Centre with approximately 40 patients/year receiving this treatment. Surgery is normally performed 6-8 weeks following chemoradiation.

One of the unique opportunities with rectal cancer is that histopathological analysis of the resected specimen gives an accurate assessment of the response of the tumour to pre-operative chemoradiation. Those tumours having a pathologic complete response (pCR) are known to have an excellent long-term outcome.

Data from the department of Radiation Oncology at Northern Sydney Cancer Centre demonstrates that only 17% of patients experience a pCR, which is in keeping with other series [4]. In this study of 48 patients, 38% showed evidence of tumour shrinkage 2 weeks into treatment as demonstrated on a Cone Beam CT scan taken during radiotherapy. 44% of these patients demonstrated a pCR following surgery. In the remaining 62%, none of these patients had a pCR. This confirms the principle that early response during radiation based only on gross tumour shrinkage can be a powerful predictor of subsequent response.

Assessing response during Cone Beam CT is very subjective and not possible on all patients. Furthermore it may be possible to better predict those likely to have a complete response with alternate imaging modalities. There is emerging data that functional and microstructural imaging modalities can also be used to predict and assess treatment response prior to, during and following the delivery of pre-operative chemoradiation for locally advanced rectal cancer [5].

Diffusion-weighted magnetic resonance imaging (DWI) is a microstructural imaging technique that characterizes tissue based on differences in the movement of water molecules. These differences can be quantified using the apparent diffusion coefficient (ADC). The ADC has been shown to differentiate post-treatment inflammation and necrosis from recurrent or persistent tumoural tissue in rectal cancer with high specificity [6]. Pre-treatment ADC assessment with DWI has a sensitivity of 100% and specificity of 86% for detection of pCR (p=0.003) whilst treatment induced changes in ADC (measured at week 2 chemoradiation) have a sensitivity and specificity of 100% for predicting pCR (p=0.0006) [5]. Prediction of response should be improved further with the use of PET imaging. A staging 18F-FDG (18F-2-fluoro-2-deoxy-D-glucose fluorodeoxyglucose ) PET/CT has been demonstrated to provide new findings compared to contrast-enhanced CT of the thorax/abdomen/pelvis resulting in a change in the stage of disease and alteration of treatment strategy in 14% of patients [7]. F-FDG PET/CT can be more predictive of histological response and outcome than anatomic imaging alone [8]. Whilst PET Response Criteria in Solid Tumors (PERCIST) has not been shown to have predictive power on the response to neoadjuvant therapy, PET Residual Disease in Solid Tumor (PREDIST) criteria has been shown to correlate to pCR (p = 0.004) [9].