Materials and Methods: Thirteen consecutive patients who underwent pre-operative embolisation of a musculoskeletal tumour followed by surgical intervention at our institution from May 2012 to January 2016 were enrolled into the study. Patient demographics, tumour characteristics, embolisation techniques and type of surgery were recorded. Technical success of embolisation, amount of blood loss during surgery and transfusion requirements were estimated.
Results: There were five female and eight male patients who underwent pre-operative embolisation during the study period. The age ranged between 16 to 68 years, and the median age was 54. Technical success was achieved in all patients. Mean intra-operative blood loss was 1403ml, with a range of 150ml to 6900ml. Eight patients (62%) required intra-operative blood products of packed red blood cells and fresh frozen plasma. No major complications occurred during embolisation.
Conclusion: Pre-operative trans-arterial embolisation is feasible and safe for a variety of large and hypervascular musculoskeletal tumours. Our small series suggests that preoperative embolisation could contribute to the reduction of the intra-operative and post-operative blood product transfusion. It should be considered as a pre-operative adjunct for major tumour resections with a high risk of bleeding. The use of the haemoglobin gap complemented the assessment of perioperative blood loss.
METHOD: In the present study, three separate data cohorts containing 1288 breast lesions from three countries (Malaysia, Iran, and Turkey) were utilized for MLmodel development and external validation. The model was trained on ultrasound images of 725 breast lesions, and validation was done separately on the remaining data. An expert radiologist and a radiology resident classified the lesions based on the BI-RADS lexicon. Thirteen morphometric features were selected from a contour of the lesion and underwent a three-step feature selection process. Five features were chosen to be fed into the model separately and combined with the imaging signs mentioned in the BI-RADS reference guide. A support vector classifier was trained and optimized.
RESULTS: The diagnostic profile of the model with various input data was compared to the expert radiologist and radiology resident. The agreement of each approach with histopathologic specimens was also determined. Based on BI-RADS and morphometric features, the model achieved an area under the receiver operating characteristic (ROC) curve (AUC) of 0.885, which is higher than the expert radiologist and radiology resident performances with AUC of 0.814 and 0.632, respectively in all cohorts. DeLong's test also showed that the AUC of the ML protocol was significantly different from that of the expert radiologist (ΔAUCs = 0.071, 95%CI: (0.056, 0.086), P = 0.005).
CONCLUSIONS: These results support the possible role of morphometric features in enhancing the already well-excepted classification schemes.