The purpose of this study was to develop a computer-assisted diagnosis (CAD) system based on image-processing techniques and artificial neural networks for the interpretation of bone scans performed to determine the presence or absence of metastases. Methods: A training group of 810 consecutive patients who had undergone bone scintigraphy due to suspected metastatic disease were included in the study. Whole-body images, anterior and posterior views, were obtained after an injection of ^99mTc-methylene diphosphonate. The image-processing techniques included algorithms for automatic segmentation of the skeleton and automatic detection and feature extraction of hot spots. Two sets of artificial neural networks were used to classify the images, 1 classifying each hot spot separately and the other classifying the whole bone scan. A test group of 59 patients with breast or prostate cancer was used to evaluate the CAD system. The patients in the test group were selected to reflect the spectrum of pathology found in everyday clinical work. As the gold standard for the test group, we used the final clinical assessment of each case. This assessment was based on follow-up scans and other clinical data, including the results of laboratory tests, and available diagnostic images, such as from MRI, CT, and radiography, from a mean follow-up period of 4.8 y. Results: The CAD system correctly identified 19 of the 21 patients with metastases in the test group, showing a sensitivity of 90%. False-positive classification of metastases was made in 4 of the 38 patients not classified as having metastases by the gold standard, resulting in a specificity of 89%. Conclusion: A completely automated CAD system can be used to detect metastases in bone scans. Application of the method as a clinical decision support tool appears to have significant potential.

The aim of this multicenter study was to investigate whether a computer-assisted diagnosis (CAD) system could improve performance and reduce interobserver variation in bone-scan interpretations of the presence or absence of bone metastases. Methods: The whole-body bone scans (anterior and posterior views) of 59 patients with breast or prostate cancer who had undergone scintigraphy for suspected bone metastatic disease were studied. The patients were selected to reflect the spectrum of pathology found in everyday clinical work. Thirty-five physicians working at 18 of the 30 nuclear medicine departments in Sweden agreed to participate. The physicians were asked to classify each case for the presence or absence of bone metastasis, without (baseline) and with the aid of the CAD system (1 y later), using a 4-point scale. The final clinical assessments, based on follow-up scans and other clinical data including the results of laboratory tests and available diagnostic images (such as MRI, CT, and radiographs from a mean follow-up period of 4.8 y), were used as the gold standard. Each physician's classification was pairwise compared with the classifications made by all the other physicians, resulting in 595 pairs of comparisons, both at baseline and after using the CAD system. Results: The physicians increased their sensitivity from 78% without to 88% with the aid of the CAD system (P < 0.001). The specificity did not change significantly with CAD. Percentage agreement and κ-values between paired physicians on average increased from 64% to 70% and from 0.48 to 0.55, respectively, with the CAD system. Conclusion: A CAD system improved physicians' sensitivity in detecting metastases and reduced interobserver variation in planar whole-body bone scans. The CAD system appears to have significant potential in assisting physicians in their clinical routine.

To develop a completely automated method, based on image processing techniques and artificial neural networks, for the interpretation of bone scans regarding the presence or absence of metastases. A total of 200 patients, all of whom had the diagnosis of breast or prostate cancer and had undergone bone scintigraphy, were studied retrospectively. Whole-body images, anterior and posterior, were obtained after injection of 99mTc-methylene diphosphonate. The study material was randomly divided into a training group and a test group, with 100 patients in each group. The training group was used in the process of developing the image analysis techniques and to train the artificial neural networks. The test group was used to evaluate the automated method. The image processing techniques included algorithms for segmentation of the head, chest, spine, pelvis and bladder, automatic thresholding and detection of hot spots. Fourteen features from each examination were used as input to artificial neural networks trained to classify the images. The interpretations by an experienced physician were used as the ‘gold standard’. The automated method correctly identified 28 of the 31 patients with metastases in the test group, i.e., a sensitivity of 90%. A false positive classification of metastases was made in 18 of the 69 patients not classified as having metastases by the experienced physician, resulting in a specificity of 74%. A completely automated method can be used to detect metastases in bone scans. Future developments in this field may lead to clinically valuable decision-support tools.

To develop an artificial intelligence (AI)-based method for the detection of focal skeleton/bone marrow uptake (BMU) in patients with Hodgkin’s lymphoma (HL) undergoing staging with FDG-PET/CT. The results of the AI in a separate test group were compared to the interpretations of independent physicians. The skeleton and bone marrow were segmented using a convolutional neural network. The training of AI was based on 153 un-treated patients. Bone uptake significantly higher than the mean BMU was marked as abnormal, and an index, based on the total squared abnormal uptake, was computed to identify the focal uptake. Patients with an index above a predefined threshold were interpreted as having focal uptake. As the test group, 48 un-treated patients who had undergone a staging FDG-PET/CT between 2017–2018 with biopsy-proven HL were retrospectively included. Ten physicians classified the 48 cases regarding focal skeleton/BMU. The majority of the physicians agreed with the AI in 39/48 cases (81%) regarding focal skeleton/bone marrow involvement. Inter-observer agreement between the physicians was moderate, Kappa 0.51 (range 0.25–0.80). An AI-based method can be developed to highlight suspicious focal skeleton/BMU in HL patients staged with FDG-PET/CT. Inter-observer agreement regarding focal BMU is moderate among nuclear medicine physicians.