ABSTRACT
Background: Structured reporting has been demonstrated to increase report completeness and to reduce error rate, also enabling data mining of radiological reports. Still, structured reporting is perceived by radiologists as a fragmented reporting style, limiting their freedom of expression. Purpose: A deep learning-based natural language processing method was developed to automatically convert unstructured COVID-19 chest CT reports into structured reports. Methods: Two hundred-two COVID-19 chest CT were retrospectively reviewed by two experienced radiologists, who wrote for each exam a free-form text radiological report and coherently filled the template provided by the Italian Society of Medical and Interventional Radiology, used as ground-truth. A semi-supervised convolutional neural network was implemented to extract 62 categorical variables from the report. Two iterations were carried-out, the first without fine-tuning, the second one performing a fine-tuning. The performance was measured using the mean accuracy and the F1 mean score. An error analysis was performed to identify errors entirely attributable to incorrect processing of the model. Results: The algorithm achieved a mean accuracy of 93.7% and an F1 score 93.8% in the first iteration. Most of the errors were exclusively attributable to wrong inference (46%). In the second iteration the model achieved for both parameters 95,8% and percentage of errors attributable to wrong inference decreased to 26%. Conclusions: The convolutional neural network achieved an optimal performance in the automated conversion of free-form text into structured radiological reports, overcoming all the limitation attributed to structured reporting and finally paving the way for data mining of radiological report.
ABSTRACT
OBJECTIVE: To assess the short- and long-term thermal impact of subclinical and clinical regimens of a single, non-invasive uniform ultrasound treatment session on subcutaneous adipose tissue (SAT). STUDY DESIGN: Prospective, open-label, single-arm, split-side study. METHODS: Patients (n = 17) were subjected to uniform ultrasound treatment, delivered in a single session with the SlimME device. The device was set to one of four treatment regimens, which differed in their durations and energy fluences during the raise and maintenance phases. Up to six abdominal regions were treated, with six patients receiving a different treatment on each side of the abdomen. Safety was assessed by measuring skin surface temperature, evaluating expected skin responses immediately and 30 min after treatment and via patient ratings of pain and discomfort. Efficacy of raising and then maintaining SAT temperatures at 48°C, was determined by routinely measuring SAT temperatures during the treatment session and by histological analysis of samples collected 7 (n = 13) or 90 (n = 4) days after treatment. RESULTS: Trace to mild erythema was observed in up to 48% of the treated zones, which, in most cases, resolved within 30 minutes. No significant rise in mean skin surface temperature (≤26.5°C) was recorded following any of the four tested regimens. Overall, patients reported tolerability to treatment, with the highest mean pain score registered for the moderate and high intensity regimens (4.4 ± 1.5 and 4.9 ± 1.4, respectively). Mean SAT temperatures did not exceed 48.4 ± 2.5°C and were effectively maintained throughout the maintenance phase of the treatment session. Low-energy fluence led to localized fat coagulative necrotic lesions, surrounded by subacute rim of inflammation, while high-energy fluence induced fat coagulative necrosis alongside granulomatous panniculitis, which resolved within 90 days. CONCLUSION: The tested uniform ultrasound regimens elicited SAT temperature elevations, with a subsequent energy-dependent increase in degree of fat necrosis. At the same time, the unique design spared the surrounding tissue from thermal damage and was associated with minimal discomfort. Taken together, the SlimME device constitutes an effective tool for destruction of stubborn hypodermal fat deposits. Lasers Surg. Med. © 2018 The Authors. Lasers in Surgery and Medicine Published by Wiley Periodicals, Inc.
