AI in medical physics: guidelines for publication.

The Abstract is intended to provide a concise summary of the study and its scientific findings. For AI/ML applications in medical physics, a problem statement and rationale for utilizing these algorithms are necessary while highlighting the novelty of the approach. A brief numerical description of how the data are partitioned into subsets for training of the AI/ML algorithm, validation (including tuning of parameters), and independent testing of algorithm performance is required. This is to be followed by a summary of the results and statistical metrics that quantify the performance of the AI/ML algorithm.

Assessment of task-based performance from five clinical DBT systems using an anthropomorphic breast phantom.

PURPOSE: Digital breast tomosynthesis (DBT) is a limited-angle tomographic breast imaging modality that can be used for breast cancer screening in conjunction with full-field digital mammography (FFDM) or synthetic mammography (SM). Currently, there are five commercial DBT systems that have been approved by the U.S. FDA for breast cancer screening, all varying greatly in design and imaging protocol. Because the systems are different in technical specifications, there is a need for a quantitative approach for assessing them. In this study, the DBT systems are assessed using a novel methodology with an inkjet-printed anthropomorphic phantom and four alternative forced choice (4AFC) study scheme. METHOD: A breast phantom was fabricated using inkjet printing and parchment paper. The phantom contained 5-mm spiculated masses fabricated with potassium iodide (KI)-doped ink and microcalcifications (MCs) made with calcium hydroxyapatite. Images of the phantom were acquired on all five systems with DBT, FFDM, and SM modalities where available using beam settings under automatic exposure control. A 4AFC study was conducted to assess reader performance with each signal under each modality. Statistical analysis was performed on the data to determine proportion correct (PC), standard deviations, and levels of significance. RESULTS: For masses, overall detection was highest with DBT. The difference in PC was statistically significant between DBT and SM for most systems. A relationship was observed between increasing PC and greater gantry span. For MCs, performance was highest with DBT and FFDM compared to SM. The difference between PC of DBT and PC of SM was statistically significant for all manufacturers. CONCLUSIONS: This methodology represents a novel approach for evaluating systems. This study is the first of its kind to use an inkjet-printed anthropomorphic phantom with realistic signals to assess performance of clinical DBT imaging systems.

Deformable Mapping Method to Relate Lesions in Dedicated Breast CT Images to Those in Automated Breast Ultrasound and Digital Breast Tomosynthesis Images.

This work demonstrates the potential for using a deformable mapping method to register lesions between dedicated breast computed tomography (bCT) and both automated breast ultrasound (ABUS) and digital breast tomosynthesis (DBT) images (craniocaudal [CC] and mediolateral oblique [MLO] views). Two multi-modality breast phantoms with external fiducial markers attached were imaged by the three modalities. The DBT MLO view was excluded for the second phantom. The automated deformable mapping algorithm uses biomechanical modeling to determine corresponding lesions based on distances between their centers of mass (dCOM) in the deformed bCT model and the reference model (DBT or ABUS). For bCT to ABUS, the mean dCOM was 5.2 ± 2.6 mm. For bCT to DBT (CC), the mean dCOM was 5.1 ± 2.4 mm. For bCT to DBT (MLO), the mean dCOM was 4.7 ± 2.5 mm. This application could help improve a radiologist's efficiency and accuracy in breast lesion characterization, using multiple imaging modalities.

Deformable mapping using biomechanical models to relate corresponding lesions in digital breast tomosynthesis and automated breast ultrasound images.

This work investigates the application of a deformable localization/mapping method to register lesions between the digital breast tomosynthesis (DBT) craniocaudal (CC) and mediolateral oblique (MLO) views and automated breast ultrasound (ABUS) images. This method was initially validated using compressible breast phantoms. This methodology was applied to 7 patient data sets containing 9 lesions. The automated deformable mapping algorithm uses finite element modeling and analysis to determine corresponding lesions based on the distance between their centers of mass (dCOM) in the deformed DBT model and the reference ABUS model. This technique shows that location information based on external fiducial markers is helpful in the improvement of registration results. However, use of external markers are not required for deformable registration results described by this methodology. For DBT (CC view) mapped to ABUS, the mean dCOM was 14.9 ±â€¯6.8 mm based on 9 lesions using 6 markers in deformable analysis. For DBT (MLO view) mapped to ABUS, the mean dCOM was 13.7 ±â€¯6.8 mm based on 8 lesions using 6 markers in analysis. Both DBT views registered to ABUS lesions showed statistically significant improvements (p ≤ 0.05) in registration using the deformable technique in comparison to a rigid registration. Application of this methodology could help improve a radiologist's characterization and accuracy in relating corresponding lesions between DBT and ABUS image datasets, especially for cases of high breast densities and multiple masses.

Evaluation of Virtual Reality for Detection of Lung Nodules on Computed Tomography.

Virtual reality (VR) systems can offer benefits of improved ergonomics, but their resolution may currently be limited for the detection of small features. For detection of lung nodules, we compared the performance of VR versus standard picture archiving and communication system (PACS) monitor. Four radiologists and 1 novice radiologist reviewed axial computed tomography (CTs) of the thorax using standard PACS monitors (SM) and a VR system (HTC Vive, HTC). In this study, 3 radiologists evaluated axial lung-window CT images of a Lungman phantom. One radiologist and the novice radiologist reviewed axial lung-window patient CT thoracic images (32 patients). This HIPAA-compliant study was approved by the institutional review board. Detection of 227 lung nodules on patient CTs did not result in different sensitivity with SM compared with VR. Detection of 23 simulated Lungman phantom lung nodules on CT with SM resulted in statistically greater sensitivity (78.3%) than with VR (52.2%, P = .041) for 1 of 3 radiologists. The trend was similar but not significant for the other radiologists. There was no significant difference in the time spent by readers reviewing CT images with VR versus SM. These findings indicate that performance of a commercially available VR system for detection of lung nodules may be similar to traditional radiology monitors for assessment of small lung nodules on CTs of the thorax for most radiologists. These results, along with the potential of improving ergonomics for radiologists, are promising for the future development of VR in diagnostic radiology.

Deformable mapping technique to correlate lesions in digital breast tomosynthesis and automated breast ultrasound images.

PURPOSE: To develop a deformable mapping technique to match corresponding lesions between digital breast tomosynthesis (DBT) and automated breast ultrasound (ABUS) images. METHODS: External fiducial markers were attached to the surface of two CIRS multi-modality compressible breast phantoms (A and B) containing multiple simulated lesions. Both phantoms were imaged with DBT (upright positioning with cranial-caudal compression) and ABUS (supine positioning with anterior-to-chest wall compression). The lesions and markers were manually segmented by three different readers. Reader segmentation similarity and reader reproducibility were assessed using Dice similarity coefficients (DSC) and distances between centers of mass (dCOM ). For deformable mapping between the modalities each reader's segmented dataset was processed with an automated deformable mapping algorithm as follows: First, Morfeus, a finite element (FE) based multi-organ deformable image registration platform, converted segmentations into triangular surface meshes. Second, Altair HyperMesh, a FE pre-processor, created base FE models for the ABUS and DBT data sets. All deformation is performed on the DBT image data; the ABUS image sets remain fixed throughout the process. Deformation was performed on the external skin contour (DBT image set) to match the external skin contour on the ABUS set, and the locations of the external markers were used to morph the skin contours to be within a user-defined distance. Third, the base DBT-FE model was deformed with the FE analysis solver, Optistruct. Deformed DBT lesions were correlated with matching lesions in the base ABUS FE model. Performance (lesion correlation) was assessed with dCOM for all corresponding lesions and lesion overlap. Analysis was performed to determine the minimum number of external fiducial markers needed to create the desired correlation and the improvement of correlation with the use of external markers. RESULTS: Average DSC for reader similarity ranged from 0.88 to 0.91 (ABUS) and 0.57 to 0.83 (DBT). Corresponding dCOM ranged from 0.20 to 0.36 mm (ABUS) and 0.11 to 1.16 mm (DBT). Lesion correlation is maximized when all corresponding markers are within a maximum distance of 5 mm. For deformable mapping of phantom A, without the use of external markers, only two of six correlated lesions showed overlap with an average lesion dCOM of 6.8 ± 2.8 mm. With use of three external fiducial markers, five of six lesions overlapped and average dCOM improved to 4.9 ± 2.4 mm. For deformable mapping of Phantom B without external markers analysis, four lesions were correlated of seven with overlap between only one of seven lesions, and an average lesion dCOM of 9.7 ± 3.5 mm. With three external markers, all seven possible lesions were correlated with overlap between four of seven lesions. The average dCOM was 8.5 ± 4.0 mm. CONCLUSION: This work demonstrates the potential for a deformable mapping technique to relate corresponding lesions in DBT and ABUS images by showing improved lesion correspondence and reduced lesion registration errors with the use of external fiducial markers. The technique should improve radiologists' characterization of breast lesions which can reduce patient callbacks, misdiagnoses and unnecessary biopsies.

Error analysis of speed of sound reconstruction in ultrasound limited angle transmission tomography.

We have investigated limited angle transmission tomography to estimate speed of sound (SOS) distributions for breast cancer detection. That requires both accurate delineations of major tissues, in this case by segmentation of prior B-mode images, and calibration of the relative positions of the opposed transducers. Experimental sensitivity evaluation of the reconstructions with respect to segmentation and calibration errors is difficult with our current system. Therefore, parametric studies of SOS errors in our bent-ray reconstructions were simulated. They included mis-segmentation of an object of interest or a nearby object, and miscalibration of relative transducer positions in 3D. Close correspondence of reconstruction accuracy was verified in the simplest case, a cylindrical object in homogeneous background with induced segmentation and calibration inaccuracies. Simulated mis-segmentation in object size and lateral location produced maximum SOS errors of 6.3% within 10 mm diameter change and 9.1% within 5 mm shift, respectively. Modest errors in assumed transducer separation produced the maximum SOS error from miscalibrations (57.3% within 5 mm shift), still, correction of this type of error can easily be achieved in the clinic. This study should aid in designing adequate transducer mounts and calibration procedures, and in specification of B-mode image quality and segmentation algorithms for limited angle transmission tomography relying on ray tracing algorithms.

Preliminary Clinical Experience with a Combined Automated Breast Ultrasound and Digital Breast Tomosynthesis System.

We analyzed the performance of a mammographically configured, automated breast ultrasound (McABUS) scanner combined with a digital breast tomosynthesis (DBT) system. The GE Invenia ultrasound system was modified for integration with GE DBT systems. Ultrasound and DBT imaging were performed in the same mammographic compression. Our small preliminary study included 13 cases, six of whom had contained invasive cancers. From analysis of these cases, current limitations and corresponding potential improvements of the system were determined. A registration analysis was performed to compare the ease of McABUS to DBT registration for this system with that of two systems designed previously. It was observed that in comparison to data from an earlier study, the McABUS-to-DBT registration alignment errors for both this system and a previously built combined system were smaller than those for a previously built standalone McABUS system.

CT breast dose reduction with the use of breast positioning and organ-based tube current modulation.

PURPOSE: This study aimed to investigate the breast dose reduction potential of a breast-positioning (BP) technique for thoracic CT examinations with organ-based tube current modulation (OTCM). METHODS: This study included 13 female anthropomorphic computational phantoms (XCAT, age range: 27-65 y.o., weight range: 52-105.8 kg). Each phantom was modified to simulate three breast sizes in standard supine geometry. The modeled breasts were then morphed to emulate BP that constrained the majority of the breast tissue inside the 120° anterior tube current (mA) reduction zone. The OTCM mA value was modeled using a ray-tracing program, which reduced the mA to 20% in the anterior region with a corresponding increase to the posterior region. The organ doses were estimated by a validated Monte Carlo program for a typical clinical CT system (SOMATOM Definition Flash, Siemens Healthcare). The simulated organ doses and organ doses normalized by CTDIvol were used to compare three CT protocols: attenuation-based tube current modulation (ATCM), OTCM, and OTCM with BP (OTCMBP ). RESULTS: On average, compared to ATCM, OTCM reduced breast dose by 19.3 ± 4.5%, whereas OTCMBP reduced breast dose by 38.6 ± 8.1% (an additional 23.8 ± 9.4%). The dose saving of OTCMBP was more significant for larger breasts (on average 33, 38, and 44% reduction for 0.5, 1, and 2 kg breasts, respectively). Compared to ATCM, OTCMBP also reduced thymus and heart dose by 15.1 ± 7.4% and 15.9 ± 6.2% respectively. CONCLUSIONS: In thoracic CT examinations, OTCM with a breast-positioning technique can markedly reduce unnecessary exposure to radiosensitive organs in anterior chest wall, specifically breast tissue. The breast dose reduction is more notable for women with larger breasts.

Fast Variance Prediction for Iteratively Reconstructed CT Images With Locally Quadratic Regularization.

Predicting noise properties of iteratively reconstructed CT images is useful for analyzing reconstruction methods; for example, local noise power spectrum (NPS) predictions may be used to quantify the detectability of an image feature, to design regularization methods, or to determine dynamic tube current adjustment during a CT scan. This paper presents a method for fast prediction of reconstructed image variance and local NPS for statistical reconstruction methods using quadratic or locally quadratic regularization. Previous methods either require impractical computation times to generate an approximate map of the variance of each reconstructed voxel, or are restricted to specific CT geometries. Our method can produce a variance map of the entire image, for locally shift-invariant CT geometries with sufficiently fine angular sampling, using a computation time comparable to a single back-projection. The method requires only the projection data to be used in the reconstruction, not a reconstruction itself, and is reasonably accurate except near image edges where edge-preserving regularization behaves highly nonlinearly. We evaluate the accuracy of our method using reconstructions of both simulated CT data and real CT scans of a thorax phantom.

Dual-Energy CT-Based Differentiation of Benign Posttreatment Changes From Primary or Recurrent Malignancy of the Head and Neck: Comparison of Spectral Hounsfield Units at 40 and 70 keV and Iodine Concentration.

OBJECTIVE: The goals of our study were to evaluate dual-energy CT (DECT) differences between benign posttreatment changes and primary or recurrent head and neck malignancies in terms of spectral Hounsfield units for virtual monochromatic series at 40 keV and iodine concentration and compare their utility with that of spectral Hounsfield units at 70 keV. MATERIALS AND METHODS: A retrospective review of patients with a history of head and neck malignancy evaluated with DECT of the neck from November 2012 through December 2014 revealed 16 patients with benign posttreatment changes and 24 with malignancies (17 primary tumors and seven recurrent tumors). One reader placed ROIs within benign posttreatment changes or malignant tumors in each patient to generate spectral Hounsfield units at 40 keV, iodine concentration, and spectral Hounsfield units at 70 keV, and the Wilcoxon rank sum test was used to evaluate the differences between the two cohorts. ROC curves were also generated, and AUC and partial AUC were calculated at the three following specificities: 75%, 80%, and 90%. RESULTS: Malignant tissues were significantly different from benign posttreatment changes in spectral Hounsfield units at 40 keV (p < 0.0001), iodine concentration (p < 0.0001), and spectral Hounsfield units at 70 keV (p = 0.0001). The AUCs were 0.949, 0.943, and 0.858 for spectral Hounsfield units at 40 keV, iodine concentration, and spectral Hounsfield units at 70 keV, respectively. Both spectral Hounsfield units at 40 keV and iodine concentration had statistically higher partial AUCs than spectral Hounsfield units at 70 keV at 90% specificity (p = 0.0133 and 0.0063, respectively) but were not significantly different from each other. CONCLUSION: DECT-derived spectral Hounsfield units at 40 keV and iodine concentration may be superior to spectral Hounsfield units at 70 keV, which is similar to MDCT, in differentiating benign posttreatment changes from primary or recurrent head and neck malignancies.

Utility and Associated Risk of Pulmonary Embolism Computed Tomography Scans in the Michigan Lupus Cohort.

OBJECTIVE: Systemic lupus erythematosus patients are frequently evaluated for chest pain and may have multiple pulmonary embolism (PE) computed tomography (CT) scans. This study was undertaken to determine the incidence of pulmonary embolism in the University of Michigan Lupus Cohort patients who have undergone PE CT scans and to estimate the associated increased risk of breast and lung cancer from radiation exposure. METHODS: We reviewed records of patients in the University of Michigan Lupus Cohort (n = 854) and determined the number and outcome of PE CT scans. Radimetrics software was used to perform individualized calculations of radiation dose to the lung and breast of each patient. We used this dose information, the patient's age at the time of scan, and risks according to the Biological Effects of Ionizing Radiation, report VII, to estimate the increased incidence risks of breast and lung cancer. RESULTS: A total of 182 of 856 patients (21%) underwent 357 PE CT scans. The overall rate of positivity was 7.5%. For patients undergoing their first through third scans, the rate of positivity for PE was 8.8%, whereas patients undergoing their fourth through tenth scans had 1.6% positivity. The highest increase in incidence risk was 0.87% for breast and 0.62% for lung. CONCLUSION: Patients with multiple previous PE CT scans had lower likelihood of a positive result on subsequent scans and higher risks of malignancy. The magnitude of risk should not discourage performance of PE CT when clinically indicated.

Surgical biopsy is still necessary for BI-RADS 4 calcifications found on digital mammography that are technically too faint for stereotactic core biopsy.

The purpose of this study was to evaluate the outcome of faint BI-RADS 4 calcifications detected with digital mammography that were not amenable to stereotactic core biopsy due to suboptimal visualization. Following Institutional Review Board approval, a HIPAA compliant retrospective search identified 665 wire-localized surgical excisions of calcifications in 606 patients between 2007 and 2010. We included all patients that had surgical excision for initial diagnostic biopsy due to poor calcification visualization, whose current imaging was entirely digital and performed at our institution and who did not have a diagnosis of breast cancer within the prior 2 years. The final study population consisted of 20 wire-localized surgical biopsies in 19 patients performed instead of stereotactic core biopsy due to poor visibility of faint calcifications. Of the 20 biopsies, 4 (20% confidence intervals 2, 38%) were malignant, 5 (25%) showed atypia and 11 (55%) were benign. Of the malignant cases, two were invasive ductal carcinoma (2 and 1.5 mm), one was intermediate grade DCIS and one was low-grade DCIS. Malignant calcifications ranged from 3 to 12 mm. The breast density was scattered in 6/19 (32%), heterogeneously dense in 11/19 (58%) and extremely dense in 2/19 (10%). Digital mammography-detected faint calcifications that were not amenable to stereotactic biopsy due to suboptimal visualization had a risk of malignancy of 20%. While infrequent, these calcifications should continue to be considered suspicious and surgical biopsy recommended.