A human-interpretable machine learning pipeline based on ultrasound to support leiomyosarcoma diagnosis.

The preoperative evaluation of myometrial tumors is essential to avoid delayed treatment and to establish the appropriate surgical approach. Specifically, the differential diagnosis of leiomyosarcoma (LMS) is particularly challenging due to the overlapping of clinical, laboratory and ultrasound features between fibroids and LMS. In this work, we present a human-interpretable machine learning (ML) pipeline to support the preoperative differential diagnosis of LMS from leiomyomas, based on both clinical data and gynecological ultrasound assessment of 68 patients (8 with LMS diagnosis). The pipeline provides the following novel contributions: (i) end-users have been involved both in the definition of the ML tasks and in the evaluation of the overall approach; (ii) clinical specialists get a full understanding of both the decision-making mechanisms of the ML algorithms and the impact of the features on each automatic decision. Moreover, the proposed pipeline addresses some of the problems concerning both the imbalance of the two classes by analyzing and selecting the best combination of the synthetic oversampling strategy of the minority class and the classification algorithm among different choices, and the explainability of the features at global and local levels. The results show very high performance of the best strategy (AUC = 0.99, F1 = 0.87) and the strong and stable impact of two ultrasound-based features (i.e., tumor borders and consistency of the lesions). Furthermore, the SHAP algorithm was exploited to quantify the impact of the features at the local level and a specific module was developed to provide a template-based natural language (NL) translation of the explanations for enhancing their interpretability and fostering the use of ML in the clinical setting.

Overlap of Suspicious and Non-Suspicious Features in the Ultrasound Evaluations of Leiomyosarcoma: A Single-Center Experience.

Leiomyosarcoma (LMS) is a rare type of mesenchymal tumor. Suspecting LMS before surgery is crucial for proper patient management. Ultrasound is the primary method for assessing myometrial lesions. The overlapping of clinical, laboratory, as well as ultrasound features between fibroids and LMS makes differential diagnosis difficult. We report our single-center experience in ultrasound imaging assessment of LMS patients, highlighting that misleading findings such as shadowing and absent or minimal vascularization may also occur in LMS. To avoid mistakes, a comprehensive evaluation of potentially overlapping ultrasound features is necessary in preoperative ultrasound evaluations of all myometrial tumors.

Machine learning applied to MRI evaluation for the detection of lymph node metastasis in patients with locally advanced cervical cancer treated with neoadjuvant chemotherapy.

PURPOSE: Concurrent cisplatin-based chemotherapy and radiotherapy (CCRT) plus brachytherapy is the standard treatment for locally advanced cervical cancer (LACC). Platinum-based neoadjuvant chemotherapy (NACT) followed by radical hysterectomy is an alternative for patients with stage IB2-IIB disease. Therefore, the correct pre-treatment staging is essential to the proper management of this disease. Pelvic magnetic resonance imaging (MRI) is the gold standard examination but studies about MRI accuracy in the detection of lymph node metastasis (LNM) in LACC patients show conflicting data. Machine learning (ML) is emerging as a promising tool for unraveling complex non-linear relationships between patient attributes that cannot be solved by traditional statistical methods. Here we investigated whether ML might improve the accuracy of MRI in the detection of LNM in LACC patients. METHODS: We analyzed retrospectively LACC patients who underwent NACT and radical hysterectomy from 2015 to 2020. Demographic, clinical and MRI characteristics before and after NACT were collected, as well as information about post-surgery histopathology. Random features elimination wrapper was used to determine an attribute core set. A ML algorithm, namely Extreme Gradient Boosting (XGBoost) was trained and validated with tenfold cross-validation. The performances of the algorithm were assessed. RESULTS: Our analysis included n.92 patients. FIGO stage was IB2 in n.4/92 (4.3%), IB3 in n.42/92 (45%), IIA1 in n.1/92 (1.1%), IIA2 in n.16/92 (17.4%) and IIB in n.29/92 (31.5%). Despite detected neither at pre-treatment and post-treatment MRI in any patients, LNM occurred in n.16/92 (17%) patients. The attribute core set used to train ML algorithms included grading, histotypes, age, parity, largest diameter of lesion at either pre- and post-treatment MRI, presence/absence of fornix infiltration at pre-treatment MRI and FIGO stage. XGBoost showed a good performance (accuracy 89%, precision 83%, recall 78%, AUROC 0.79). CONCLUSIONS: We developed an accurate model to predict LNM in LACC patients in NACT, based on a ML algorithm requiring few easy-to-collect attributes.

A retrospective study of preoperative CA 125 levels in 82 patients with ovarian cancer.

INTRODUCTION: The aim of our study was to investigate preoperative serum CA 125 as a prognostic factor in patients with ovarian carcinoma. METHODS: A retrospective analysis was conducted on 82 patients with ovarian carcinoma treated at our Unit between 1998 and 2000 who had a serum CA 125, evaluated by a commercially available radioimmunoassay, prior to cytoreductive surgery. We looked for an association between preoperative CA 125 and known prognostic factors of ovarian cancer. We compared outcomes of patients with preoperative CA 125 at or below to 500 U/ml with outcomes of patients with preoperative CA 125 above 500 U/ml. RESULTS: A significant ( p<0.002) correlation between stage and CA 125 serum levels was found as 16 out of 18 stage I-II patients (89%) had CA 125 level 500 U/ml. Among stage III and IV patients there was nonstatistically significant relation between serum CA 125 and histologic grade (G1+G2 vs. G3) and residual disease (<1 cm vs. >1 cm) after primary cytoreductive surgery. Preoperative serum CA-125 level did not predict either recurrences or disease free interval. CONCLUSION: Preoperative CA 125 correlated well with FIGO stage but not with age, grade, residual disease after primary surgery, relapse and disease free interval.

Skin metastases in ovarian carcinoma: a report of nine cases and a review of the literature.

OBJECTIVE: Cutaneous involvement is unusual at presentation and during the course of ovarian carcinoma. The aim of the present study was to determine the incidence, clinicopathologic characteristics and prognostic factors of skin metastases in ovarian cancer patients. METHODS: A retrospective chart review was conducted on 220 patients with epithelial ovarian carcinoma treated at our unit between 1991 and 2001. Pertinent clinical information, pathologic data, treatment, and prognostic factors for survival following documentation of skin metastases were collected. Survival time was calculated from the time of diagnosis of ovarian cancer and from the time of diagnosis of the cutaneous metastasis. RESULTS: FIGO stage at the time of ovarian cancer diagnosis was stage III = five patients (56%), and stage I and stage IV = two patients each (22%). Most patients had serous papillary cystoadenocarcinoma of the ovary (78%), and one each had endometrioid and mucinous carcinomas (12%). Seven patients (78%) had poorly differentiated tumors. Only one patient had a skin metastasis at the time of ovarian cancer diagnosis; in the remaining patients the average time of appearance of skin metastases after the diagnosis of ovarian cancer was 23.4 +/- 12 months (range 4 to 37). The diameter of the skin lesions ranged between 0.5 and 3 cm. Three patients had a single skin nodule, while six had multiple skin lesions. Eight patients (89%) have died of disease and median survival after diagnosis of the skin metastases was 4 months (range 2 to 65). One patient (Cases 1) is alive without tumor 4 months after diagnosis of the skin metastases. Overall survival after diagnosis of skin metastasis from ovarian cancer was 4 months (range 2 to 65). CONCLUSION: Skin involvement is a late complication that occurs rarely in ovarian cancer patients. Prognosis after skin metastases is poor and the most important prognostic factor associated with survival is the interval time between diagnosis of ovarian cancer and documentation of cutaneous involvement.