Pulmonary nodules and metastases in colorectal cancer

Patients with newly diagnosed colorectal cancer (CRC) are subjected to a preoperative thoraco-abdominal CT scan to determine the cancer stage. This staging is of relevance with regard to treatment and prognosis. About 20% of the patients have distant metastatic spread at the time of diagnosis, i.e. synchronous metastases. Most common are hepatic metastases followed by pulmonary involvement. The optimal staging modality for detecting synchronous pulmonary metastases is debated. It has been argued, that synchronous pulmonary metastases (SPCM) are rare in CRC and that the consequence of detecting SPCM is minimal. Furthermore, the current staging practice is complicated by a high number of incidental findings on the thoracic CT, so-called indeterminate pulmonary nodules (IPN). IPN can potentially represent SPCM. The purpose of this thesis was to estimate the prevalence, characteristics and clinical significance of IPN and SPCM detected at the primary staging in CRC. Study I was a systematic review of published studies on IPN in CRC focusing on the prevalence and radiological characteristics of IPN proving to be malignant. This knowledge would be of value in management strategies for IPN. On average 9% of all patients staged with a thoracic CT had IPN, however, the prevalence varied significantly between patients series. This was mainly attributed to varying/lacking definitions on IPN and variable radiological expertise in the assessment of the scans. Data were too inconsistently reported in the case series for a robust statement to be made on potential radiological characteristics suggestive of malignancy in IPN. Lymph node metastasis was the most common clinicopathological finding associated with malignancy of IPN. In conclusion, one patient of every 100 scanned patients had an IPN proving to a SPCM at follow-up, but we found no evidence that IPN should result in intensified diagnostic work-up besides routine follow-up for CRC. Study II was an analysis of the prevalence of and risk factors for SPCM in a Danish nationwide cohort of patients with newly diagnosed CRC from 2001 to 2011. SPCM were detected in 7.5% of the patients and in 37% of these cases the metastatic spread was confined to the lungs. The prevalence of SPCM increased with the implementation of thoracic CT in CRC staging. SPCM impaired survival significantly and was associated with increasing age and rectal cancer. Resection of the primary tumour, resection of the SPCM and treatment with chemotherapy were associated with improved survival. Unfortunately, however, only very few patients were subjected to all three treatment measures, and the improved prognosis could simply be the result of a selection bias. The inter-observer variation in classification of findings at thoracic CT scans was investigated in study III and was based on staging CT scans from a local cohort of patients with newly diagnosed CRC. Based on an experienced thoracic radiologist's assessment of the scans, we searched for radiological characteristics of IPN that could predict malignancy of the nodule. There was a significant difference in the number of IPNs detected between the primary and the thoracic radiologist's assessment. The thoracic radiologist classified fewer nodules as IPN and even reported with higher specificity and sensitivity for SPCM. More than 20% of IPNs (as classified by the thoracic radiologist) proved to be SPCM. Unfortunately, no radiological characteristics could be associated with a malignant outcome. In continuation of these findings we suggested a secondary review of scans with IPN be a group of dedicated thoracic radiologists. This approach might reduce the need for additional work-up for IPN and calls for clarification in future prospective studies. Identification of patients in particular risk of SPCM could be of value in the assessment of pulmonary nodules. Several biomarkers have been proposed for differential metastatic patterns in CRC. In study IV we investigated the significance of mismatch repair status (MMR) for the location of metastatic spread in a nationwide Danish cohort of patients with newly diagnosed CRC between 2010 and 2012. Deficient MMR was associated with a reduced risk of synchronous metastatic disease, however, the risk reduction only applied to hepatic metastases. MMR had no impact on SPCM and is therefore currently of no use in the assessment of IPN.