The differential diagnoses of a retroperitoneal sarcoma are diverse. Before considering a diagnosis of sarcoma, meticulous clinico-pathological effort should be made to exclude the possibility of sarcomatoid carcinoma or mesothelioma, metastasis, and benign pseudo-sarcomatous entities. Careful correlation with the clinician and radiologist may yield critical information about the tumour (e.g. history of familial syndromes, necrosis, relationship with adjacent organs) that will greatly aid in narrowing the differential diagnoses.
In adults, the most common groups of retroperitoneal sarcomas are often high-grade, which includes liposarcomas, leiomyosarcomas and undifferentiated pleomorphic sarcoma.7Liposarcomas can be broadly divided into three unique classes: well-differentiated/de-differentiated, myxoid/round cell, and pleomorphic subtypes. Retroperitoneal sarcomas in paediatric patients are uncommon and more likely to be rhabomyosarcomas, extraosseous primitive neuroectodermal tumour, neuroblastomas, and Wilms tumours.8,9
If biopsy specimens are required, preliminary investigation generally involves obtaining multiple core biopsies (at least five cores) using needles of at least 16 gauge to subtype the tumour and assess tissue viability for further ancillary testing. Malignancy grading may be underestimated in small biopsies; therefore, a final diagnosis can only be reliably obtained after definitive tumour resection. Frozen-section technique for immediate diagnosis should be avoided due to suboptimal histological interpretation and lack of immediate ancillary support.
As with sarcoma elsewhere, tumour site, size, depth, vascular invasion and margin status should be recorded. Detailed discussions of the tumour datasets and gross handling are comprehensively covered in various pathology colleges and organisations guidelines.10-12For retroperitoneal sarcomas where tumours are often large and voluminous, a few practical suggestions could be made. Firstly, coordination between the surgeon and pathologist is important to ensure that fresh tissue is available for molecular studies and tumour banking. Secondly, intelligent sampling of critical sites (e.g. interface region in a dedifferentiated tumour) rather than blind extensive sampling is more cost-effective in obtaining the correct diagnosis. Thirdly, tumours with heterogeneous appearance should be generously sampled to investigate heterologous or dedifferentiated elements.