Utility of Diffusion-weighted MR Imaging in Guiding Selective Percutaneous Drainage of Postoperative Intra-abdominal Abscesses After Colorectal Resection

Percutaneous drainage of intra-abdominal abscesses occurring as a complication of colon and rectal resection has been a major advance in the management of surgical patients. Proper patient selection is critical for safe and effective management in this population. Almost ¾ of patients undergoing CT scan after colorectal resection due to clinical suspicion of intrabdominal process will have at least one fluid collection identified. These collections can represent a spectrum of clinical entities and there is not a consensus on the most effective management of these collections or even the definition of abscess. Currently, reliance on radiologic criteria in isolation can lead to overuse of interventional procedures. For example, 40% of rim-enhancing collections are sterile on aspiration. The surgeons' clinical suspicion for abscess and radiologic proximity to an anastomosis are the only criteria that are useful in predicting abscess versus sterile collection. A further consideration is the natural history of these abscesses. Studies in the diverticulitis literature have demonstrated that abscesses less than 3 cm in greatest dimension are successfully managed with antibiotics alone, while abscesses greater than 6.5 cm are likely to require intervention. However, this leaves a great number of abscesses between 3 cm and 6.5 cm that fall into uncertain grounds. In contrast to diverticulitis, where it can be reasonably inferred that an associated abdominopelvic collection is indeed and abscess, management of fluid collections identified post-operatively and determination of who will benefit from drainage is less clear. A novel radiologic technique with high discrimination between sterile and infected collection would be of great clinical utility in the post-operative management of fluid collections after colorectal resection.

The proposed research project seeks to broaden applicability of a proven but rarely used method, diffusion weighted magnetic resonance imaging (DW-MRI), to discriminate sterile or benign from infected abdominopelvic fluid collections, in order to enhance the utilization of percutaneous drainage in the post-operative setting after colorectal resection. Ability to streamline a limited MRI protocol to efficiently obtain diffusion weighted imaging of the abdominal cavity will be key to applying this methodology into daily practice. Secondly, it is not known whether DW-MRI can effectively discriminate specifically post-operative sterile collections from abscesses. This is one of the primary aims of this pilot study and will be used to generate hypothesis for a full-scale study.

Adult patients who have undergone a colon or rectal resection at the University of Chicago Medicine and have developed a CT-proven abdominopelvic fluid collection > 3 cm in greatest dimension will be eligible for inclusion in the study. The surgeon will then determine using traditional clinical and radiologic factors whether percutaneous drainage is desired. All patients with discrete abdominopelvic fluid collections > 3 cm in greatest dimension will undergo DW-MRI, and then be taken directly to the interventional radiology suite for drainage, or continue best medical care if drainage is felt to be unnecessary. Fluid will be sent for culture and gram stain, as well as cell count. In addition, the interventional radiology team will qualify the fluid in their notation as 'purulent', 'serous', 'sanguinous', etc.. Any patients initially managed without drainage who fail this medical management and ultimately undergo percutaneous drainage will be considered the crossover group.

DW-MRI readings will be analyzed and ADC values will be compared and validated in an attempt to report a threshold apparent diffusion coefficient (ADC) that reliably discriminates sterile and infected post-operative fluid collections. The effect size from utilization of DW-MRI (defined as a change in management decision to pursue drainage or hold on drainage) to determine if this is a useful clinical tool. Because it is not yet know how DW-MRI should be interpreted in this clinical setting, the DW-MRI acquisition will not be utilized in any way in the clinical care of the study patients. The DW-MRI data is considered a single-point intervention and will only be used to define an ADC threshold that can discriminate sterile from infected collections. The data will help us to identify potential limitations of DWI in differentiation of abscess from non-infected collections and allow us to determine the appropriate cohort size for a future clinical trial. Based on the correlation with DWI and clinical data, we will explore and suggest novel acquisition and analysis methods for quantitative DWI.