Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept.

[18F]Fluorodeoxyglucose positron emission tomography (FDG-PET) and computed tomography (CT) are indispensable components in modern medicine. Although PET can provide additional diagnostic value, it is costly and not universally accessible, particularly in low-income countries. To bridge this gap, we have developed a conditional generative adversarial network pipeline that can produce FDG-PET from diagnostic CT scans based on multi-center multi-modal lung cancer datasets (n = 1,478). Synthetic PET images are validated across imaging, biological, and clinical aspects. Radiologists confirm comparable imaging quality and tumor contrast between synthetic and actual PET scans. Radiogenomics analysis further proves that the dysregulated cancer hallmark pathways of synthetic PET are consistent with actual PET. We also demonstrate the clinical values of synthetic PET in improving lung cancer diagnosis, staging, risk prediction, and prognosis. Taken together, this proof-of-concept study testifies to the feasibility of applying deep learning to obtain high-fidelity PET translated from CT.

Multimodality imaging of mediastinal masses and mimics.

A wide variety of neoplastic and nonneoplastic conditions occur in the mediastinum. Imaging plays a central role in the evaluation of mediastinal pathologies and their mimics. Localization of a mediastinal lesion to a compartment and characterization of morphology, density/signal intensity, enhancement, and mass effect on neighboring structures can help narrow the differentials. The International Thymic Malignancy Interest Group (ITMIG) established a cross-sectional imaging-derived and anatomy-based classification system for mediastinal compartments, comprising the prevascular (anterior), visceral (middle), and paravertebral (posterior) compartments. Cross-sectional imaging is integral in the evaluation of mediastinal lesions. Computed tomography (CT) and magnetic resonance imaging (MRI) are useful to characterize mediastinal lesions detected on radiography. Advantages of CT include its widespread availability, fast acquisition time, relatively low cost, and ability to detect calcium. Advantages of MRI include the lack of radiation exposure, superior soft tissue contrast resolution to detect invasion of the mass across tissue planes, including the chest wall and diaphragm, involvement of neurovascular structures, and the potential for dynamic sequences during free-breathing or cinematic cardiac gating to assess motion of the mass relative to adjacent structures. MRI is superior to CT in the differentiation of cystic from solid lesions and in the detection of fat to differentiate thymic hyperplasia from thymic malignancy.

Imaging of Malignant Pleural, Pericardial, and Peritoneal Mesothelioma.

Malignant mesothelioma is a rare tumor arising from the mesothelial cells that line the pleura, pericardium, peritoneum, and tunica vaginalis. Imaging plays a primary role in the diagnosis, staging, and management of malignant mesothelioma. Multimodality imaging, including radiography, computed tomography (CT), magnetic resonance imaging (MRI), and F-18 fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT), is used in a variety of scenarios, including diagnosis, guidance for tissue sampling, staging, and reassessment of disease after therapy. CT is the primary imaging modality used in staging. MRI has superior contrast resolution compared with CT and can add value in terms of determining surgical resectability in equivocal cases. MRI can further assess the degree of local invasion, particularly into the mediastinum, chest wall, and diaphragm, for malignant pleural and pericardial mesotheliomas. FDG PET/CT plays a role in the diagnosis and staging of malignant pleural mesothelioma (MPM) and has been shown to be more accurate than CT, MRI, and PET alone in the staging of malignant pleural mesothelioma. PET/CT can also be used to target lesions for biopsy and to assess prognosis, treatment response, and tumor recurrence.

Pearls and pitfalls in imaging of mediastinal masses.

In imaging of the mediastinum, advances in computed tomography (CT), and magnetic resonance imaging (MRI) technology enable improved characterization of mediastinal masses. Knowledge of the boundaries of the mediastinal compartments is key to accurate localization. Awareness of distinguishing imaging characteristics allows radiologists to suggest a specific diagnosis or narrow the differential. In certain situations, MRI adds value to further characterize mediastinal lesions.

Habitat Imaging Biomarkers for Diagnosis and Prognosis in Cancer Patients Infected with COVID-19.

OBJECTIVES: Cancer patients have worse outcomes from the COVID-19 infection and greater need for ventilator support and elevated mortality rates than the general population. However, previous artificial intelligence (AI) studies focused on patients without cancer to develop diagnosis and severity prediction models. Little is known about how the AI models perform in cancer patients. In this study, we aim to develop a computational framework for COVID-19 diagnosis and severity prediction particularly in a cancer population and further compare it head-to-head to a general population. METHODS: We have enrolled multi-center international cohorts with 531 CT scans from 502 general patients and 420 CT scans from 414 cancer patients. In particular, the habitat imaging pipeline was developed to quantify the complex infection patterns by partitioning the whole lung regions into phenotypically different subregions. Subsequently, various machine learning models nested with feature selection were built for COVID-19 detection and severity prediction. RESULTS: These models showed almost perfect performance in COVID-19 infection diagnosis and predicting its severity during cross validation. Our analysis revealed that models built separately on the cancer population performed significantly better than those built on the general population and locked to test on the cancer population. This may be because of the significant difference among the habitat features across the two different cohorts. CONCLUSIONS: Taken together, our habitat imaging analysis as a proof-of-concept study has highlighted the unique radiologic features of cancer patients and demonstrated effectiveness of CT-based machine learning model in informing COVID-19 management in the cancer population.

Cardiac Magnetic Resonance Predicting Outcomes Among Patients at Risk for Cardiac AL Amyloidosis.

Introduction: Patients with systemic AL amyloidosis (AL) should be evaluated for cardiac amyloidosis (CA), as prognosis is strongly related to cardiac involvement. We assessed the characteristics of patients referred to cardiac magnetic resonance (CMR) with suspected CA from a cancer center and determine predictors of mortality/heart failure hospitalizations (HFH). Methods: Forty-four consecutive patients referred for CMR with suspected CA were retrospectively included. Variables collected included cardiac biomarkers, in addition to echocardiographic and CMR variables. Survival analyses were performed to determine which variables were more predictive of mortality and HFH. Results: Of the 44 patients included, 55% were females. 73% of patients were diagnosed with CA by CMR; 56% of them had an established diagnosis of AL. Patients with CA by CMR had higher native T1, higher extracellular volume (ECV) fraction, higher T2, less negative GLS by Echo, and higher troponin I and B-type natriuretic peptide (BNP). Kaplan-Meier survival analysis revealed that the following were predictive of mortality: an ECV ≥ 0.50 (p = 0.0098), CMR LVEF < 50% (p = 0.0010), T2/ECV ≤ 100 (p = 0.0001), and troponin I > 0.03 (p = 0.0025). In a stepwise conditional Cox logistic regression model, the only variable predictive of a composite of mortality and HFH was ECV (HR: 1.17, 95% CI = 1.02-1.34 p = 0.030). Conclusion: ECV seems to be an important biomarker that could be a predictor of outcomes in cardiac AL amyloidosis. In combination, CMR and serum cardiac biomarkers might help to establish prognosis in patients with CA.

Incidence and Onset of Severe Cardiac Events After Radiotherapy for Esophageal Cancer.

INTRODUCTION: Late cardiotoxicity related to radiotherapy (RT) in breast cancer and Hodgkin's lymphoma has been well-reported. However, the relatively higher cardiac dose exposure for esophageal cancer (EC) may result in the earlier onset of cardiac diseases. In this report, we examined the incidence, onset, and long-term survival outcomes of high-grade cardiac events after RT in a large cohort of patients with EC. METHODS: Between March 2005 and August 2017, a total of 479 patients with EC from a prospectively maintained institutional database at The University of Texas MD Anderson Cancer Center were analyzed. All patients were treated with either intensity-modulated RT or proton beam therapy, either preoperatively or definitively. We focused on any grade 3 or higher (G3+) cardiac events according to the Common Terminology Criteria for Adverse Events, version 5.0. RESULTS: G3+ cardiac events occurred in 18% of patients at a median of 7 months with a median follow-up time of 76 months. Preexisting cardiac disease (p = 0.001) and radiation modality (intensity-modulated RT versus proton beam therapy) (p = 0.027) were significantly associated with G3+ cardiac events. Under multivariable analysis, the mean heart dose, particularly of less than 15 Gy, was associated with reduced G3+ events. Furthermore, G3+ cardiac events were associated with worse overall survival (p = 0.041). CONCLUSIONS: Severe cardiac events were relatively common in patients with early onset EC after RT, especially those with preexisting cardiac disease and higher radiation doses to the heart. Optimal treatment approaches should be taken to reduce cumulative doses to the heart, especially for patients with preexisting cardiac disease.

Characterization of continuous bed motion effects on patient breathing and respiratory motion correction in PET/CT imaging.

Continuous bed motion (CBM) was recently introduced as an alternative to step-and-shoot (SS) mode for PET/CT data acquisition. In CBM, the patient is continuously advanced into the scanner at a preset speed, whereas in SS, the patient is imaged in overlapping bed positions. Previous investigations have shown that patients preferred CBM over SS for PET data acquisition. In this study, we investigated the effect of CBM versus SS on patient breathing and respiratory motion correction. One hundred patients referred for PET/CT were scanned using a Siemens mCT scanner. Patient respiratory waveforms were recorded using an Anzai system and analyzed using four methods: Methods 1 and 2 measured the coefficient of variation (COV) of the respiratory cycle duration (RCD) and amplitude (RCA). Method 3 measured the respiratory frequency signal prominence (RSP) and method 4 measured the width of the HDChest optimal gate (OG) window when using a 35% duty cycle. Waveform analysis was performed over the abdominothoracic region which exhibited the greatest respiratory motion and the results were compared between CBM and SS. Respiratory motion correction was assessed by comparing the ratios of SUVmax, SUVpeak, and CNR of focal FDG uptake, as well as Radiologists' visual assessment of corresponding image quality of motion corrected and uncorrected images for both acquisition modes. The respiratory waveforms analysis showed that the RCD and RCA COV were 3.7% and 33.3% lower for CBM compared to SS, respectively, while the RSP and OG were 30.5% and 2.0% higher, respectively. Image analysis on the other hand showed that SUVmax, SUVpeak, and CNR were 8.5%, 4.5%, and 3.4% higher for SS compared to CBM, respectively, while the Radiologists' visual comparison showed similar image quality between acquisition modes. However, none of the results showed statistically significant differences between SS and CBM, suggesting that motion correction is not impacted by acquisition mode.

Coronary Artery Dose-Volume Parameters Predict Risk of Calcification After Radiation Therapy.

BACKGROUND: Radiation exposure increases the risk of coronary artery disease (CAD). We explored the association of CAD with coronary artery dose-volume parameters in patients treated with 3D-planned radiation therapy (RT). METHODS: Patients who received thoracic RT and were evaluated by cardiac computed tomography ≥ 1 year later were included. Demographic data and cardiac risk factors were retrospectively collected. Dosimetric data (mean heart dose, dmax, dmean, V50 - V5) were collected for the whole heart and for each coronary artery. A coronary artery calcium (CAC) Agatston score was calculated on a per-coronary basis and as a total score. Multivariable generalized linear mixed models were generated. The predicted probabilities were used for receiver operating characteristic analyses. RESULTS: Twenty patients with a median age of 53 years at the time of RT were included. Nine patients (45%) had ≥ 3/6 conventional cardiac risk factors. Patients received RT for breast cancer (10, 50%), lung cancer (6, 30%), or lymphoma/myeloma (4, 20%) with a median dose of 60 Gy. CAC scans were performed a median of 32 months after RT. CAC score was significantly associated with radiation dose and presence of diabetes. In a multivariable model adjusted for diabetes, segmental coronary artery dosimetric parameters (dmax, dmean, V50, V40 V30, V20, V10, and V5) were significantly associated with CAC score > 0. V50 had the highest area under the ROC curve (0.89, 95% confidence interval, 0.80-0.97). CONCLUSIONS: Coronary artery radiation exposure is strongly correlated with subsequent segmental CAC score. Coronary calcification may occur soon after RT and in individuals with conventional cardiac risk factors.

MRI in cardio-oncology: A review of cardiac complications in oncologic care.

From detailed characterization of cardiac abnormalities to the assessment of cancer treatment-related cardiac dysfunction, cardiac MRI is playing a growing role in the evaluation of cardiac pathology in oncology patients. Current guidelines are now incorporating the use of MRI for the comprehensive multidisciplinary approach to cancer management, and innovative applications of MRI in research are expanding its potential to provide a powerful noninvasive tool in the arsenal against cancer. This review focuses on the application of cardiac MRI to diagnose and manage cardiovascular complications related to cancer and its treatment. Following an introduction to current cardiac MRI methods and principles, this review is divided into two sections: functional cardiovascular analysis and anatomical or tissue characterization related to cancer and cancer therapeutics. Level of Evidence: 5 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2019;50:1349-1366.

Pitfalls and Misinterpretations of Cardiac Findings on PET/CT Imaging: A Careful Look at the Heart in Oncology Patients.

Positron emission tomography (PET) computed tomography (CT) with 2-[fluorine-18] fluoro-2-deoxy-d-glucose (FDG) has been established as an effective modality for evaluation of cancer. Interpretations of patterns of physiologic 18F-FDG uptake by the heart is particularly difficult given the wide normal variations of 18F-FDG metabolic activity observed. Atypical patterns of focal or diffuse physiologic cardiac 18F-FDG uptake and post-therapeutic effects after radiation therapy, systemic diseases, or cardiomyopathy may also be confused with malignant disease on 18F-FDG PET/CT. In this article, we review the variations of normal cardiac 18F-FDG uptake observed in oncology patients and the appearances of other patterns of pathologic metabolic activity, related or not related to the malignancy being investigated, that may lead to false-negative and false-positive results.

Implications for high-precision dose radiation therapy planning or limited surgical resection after percutaneous computed tomography-guided lung nodule biopsy using a tract sealant.

PURPOSE: Precision radiation therapy such as stereotactic body radiation therapy and limited resection are being used more frequently to treat intrathoracic malignancies. Effective local control requires precise radiation target delineation or complete resection. Lung biopsy tracts (LBT) on computed tomography (CT) scans after the use of tract sealants can mimic malignant tract seeding (MTS) and it is unclear whether these LBTs should be included in the calculated tumor volume or resected. This study evaluates the incidence, appearance, evolution, and malignant seeding of LBTs. METHODS AND MATERIALS: A total of 406 lung biopsies were performed in oncology patients using a tract sealant over 19 months. Of these patients, 326 had follow-up CT scans and were included in the study group. Four thoracic radiologists retrospectively analyzed the imaging, and a pathologist examined 10 resected LBTs. RESULTS: A total of 234 of 326 biopsies (72%, including primary lung cancer [n = 98]; metastases [n = 81]; benign [n = 50]; and nondiagnostic [n = 5]) showed an LBT on CT. LBTs were identified on imaging 0 to 3 months after biopsy. LBTs were typically straight or serpiginous with a thickness of 2 to 5 mm. Most LBTs were unchanged (92%) or decreased (6.3%) over time. An increase in LBT thickness/nodularity that was suspicious for MTS occurred in 4 of 234 biopsies (1.7%). MTS only occurred after biopsy of metastases from extrathoracic malignancies, and none occurred in patients with lung cancer. CONCLUSIONS: LBTs are common on CT after lung biopsy using a tract sealant. MTS is uncommon and only occurred in patients with extrathoracic malignancies. No MTS was found in patients with primary lung cancer. Accordingly, potential alteration in planned therapy should be considered only in patients with LBTs and extrathoracic malignancies being considered for stereotactic body radiation therapy or wedge resection.

Evaluation and Management of Cardiac Tumors.

PURPOSE OF REVIEW: Our purpose is to discuss the importance of multimodality imaging in the assessment of cardiac tumors and management. We have compiled a recent review of the scientific literature and embedded our clinical pathways and recommendations based on data and clinical experience. RECENT FINDINGS: The use of contrast echocardiography in the assessment of cardiac masses has been shown to be helpful in distinguishing tumor from thrombus. Deformation imaging of cardiac tumors has been shown to differentiate better rhabdomyomas from fibromas in pediatric patients. Cardiac MRI (CMR) appears to be helpful in determining whether cardiac tumors are benign or malignant by identifying presence of infiltration, uptake of contrast in first pass perfusion and gadolinium enhancement. Patients with evidence of cardiac metastases by CMR show similar survival to stage IV cancer without cardiac metastases. In our institution, we use a standardized approach for the evaluation of cardiac masses, which includes multimodality imaging in the appropriate clinical context. The autotransplantation surgical technique has shown some promise in improving survival in patients with primary cardiac sarcomas. In our institution, we do not routinely recommend anticoagulation for "tumor-thrombus" in renal cell carcinoma due to risk of bleeding from primary tumor. Cardiac masses are often found incidentally, but sometimes can present with cardiovascular symptoms due to obstruction and valvular dysfunction, which may prompt imaging. It is important to determine whether the mass is a normal variant, imaging artifact, vegetation, thrombus, or tumor. Transthoracic echocardiography is ideally suited to be the initial imaging modality because of the portability, wide availability, lack of radiation, and relatively low cost. The gold standard cardiac imaging technique to distinguish tumor from thrombus is contrast enhanced CMR with prolonged inversion time. Advantages of CMR when compared to echocardiography regarding characterization of cardiac tumors are as follows: larger field of view, better spatial resolution, better tissue characterization, lack of attenuation, and ability to image at any prescribed plane. Primary and secondary cardiac tumors have particular characteristics in echocardiography and CMR. Imaging of cardiac tumors plays an important role in establishing a diagnosis and in planning management.

ACR-SPR-STR Practice Parameter for the Performance of Cardiac Positron Emission Tomography - Computed Tomography (PET/CT) Imaging.

This clinical practice parameter has been developed collaboratively by the American College of Radiology (ACR), the Society for Pediatric Radiology (SPR), and the Society of Thoracic Radiology (STR). This document is intended to act as a guide for physicians performing and interpreting positron emission tomography-computed tomography (PET/CT) of cardiac diseases in adults and children. The primary value of cardiac PET/CT imaging include evaluation of perfusion, function, viability, inflammation, anatomy, and risk stratification for cardiac-related events such as myocardial infarction and death. Optimum utility of cardiac PET/CT is achieved when images are interpreted in conjunction with clinical information and laboratory data. Measurement of myocardial blood flow, coronary flow reserve and detection of balanced ischemia are significant advantages of cardiac PET perfusion studies. Increasingly cardiac PET/CT is used in diagnosis and treatment response assessment for cardiac sarcoidosis.

Prospective cohort study of cancer patients diagnosed with incidental venous thromboembolism on routine computed tomography scans.

PURPOSE: Venous thromboembolism (VTE) is a major complication of cancer with recent increasing reports of incidental VTE. The objectives are to estimate the prevalence of incidental VTE in cancer patients on staging CT scans, identify common symptoms, and determine VTE recurrence in a prospective study. PATIENTS AND METHODS: One thousand ninety patients were studied. Adult cancer patients scheduled for outpatient staging CT scans were eligible. VTE cases were followed for 6 months. Fisher's exact test for group comparisons of categorical variables and generalized linear modeling to estimate the prevalence of incidental VTE was used. RESULTS: The mean age was 58 years (range 18-87 years); 50% were male. The prevalence of incidental VTE was 1.8% (CI 1.15-2.87%). Significant symptoms in patients with VTE included fatigue (p = 0.004), stress (p = 0.0195), depression (p = 0.019), poorer quality of life (p = 0.0194), and poorer physical well-being (p = 0.0007). All the patients with VTE had at least one comorbidity (p = 0.03). No patient had recurrence within 6 months. CONCLUSION: The prevalence of incidental VTE on staging CT scans is lower than previously reported. Symptoms were associated with VTE; however, further work is needed to understand whether these are clinically relevant. No VTE recurrences were noted following 6 months.

Characterizing proton-activated materials to develop PET-mediated proton range verification markers.

Conventional proton beam range verification using positron emission tomography (PET) relies on tissue activation alone and therefore requires particle therapy PET whose installation can represent a large financial burden for many centers. Previously, we showed the feasibility of developing patient implantable markers using high proton cross-section materials ((18)O, Cu, and (68)Zn) for in vivo proton range verification using conventional PET scanners. In this technical note, we characterize those materials to test their usability in more clinically relevant conditions. Two phantoms made of low-density balsa wood (~0.1 g cm(-3)) and beef (~1.0 g cm(-3)) were embedded with Cu or (68)Zn foils of several volumes (10-50 mm(3)). The metal foils were positioned at several depths in the dose fall-off region, which had been determined from our previous study. The phantoms were then irradiated with different proton doses (1-5 Gy). After irradiation, the phantoms with the embedded foils were moved to a diagnostic PET scanner and imaged. The acquired data were reconstructed with 20-40 min of scan time using various delay times (30-150 min) to determine the maximum contrast-to-noise ratio. The resultant PET/computed tomography (CT) fusion images of the activated foils were then examined and the foils' PET signal strength/visibility was scored on a 5 point scale by 13 radiologists experienced in nuclear medicine. For both phantoms, the visibility of activated foils increased in proportion to the foil volume, dose, and PET scan time. A linear model was constructed with visibility scores as the response variable and all other factors (marker material, phantom material, dose, and PET scan time) as covariates. Using the linear model, volumes of foils that provided adequate visibility (score 3) were determined for each dose and PET scan time. The foil volumes that were determined will be used as a guideline in developing practical implantable markers.

Cardiovascular involvement by osteosarcoma: an analysis of 20 patients.

BACKGROUND: Although hematogenous spread of osteosarcoma is well known, the imaging findings of cardiovascular involvement by osteosarcoma are seldom reported and can be difficult to recognize. The enhanced resolution of modern CT and MRI scanners may lead to better detection of cardiovascular involvement. OBJECTIVE: To describe the key imaging findings and clinical behavior of cardiovascular involvement by osteosarcoma. MATERIALS AND METHODS: We retrospectively reviewed the imaging findings and clinical characteristics of 20 patients with cardiovascular involvement by osteosarcoma identified by two pediatric radiologists from a review of imaging studies at our institution from 2007 to 2013. RESULTS: At initial diagnosis, the median age of the patients was 15.1 years (range 4.8-24.6 years), and 7 (35%) patients had detectable metastases. Median time to detection of cardiovascular metastases was 1.8 years (range 0-7.3 years). Sixteen patients died of disease; 4 have survived a median of 7.4 years since initial diagnosis. The sites of cardiovascular involvement were the systemic veins draining the primary and metastatic osteosarcoma, pulmonary arteries, pulmonary veins draining the pulmonary metastases, and heart. A dilated and mineralized terminal pulmonary arteriole is an early sign of metastatic osteosarcoma in the lung. Unfamiliarity with the imaging features resulted in under-recognition and misinterpretation of intravascular tumor thrombus as bland thrombus. CONCLUSION: Knowledge of imaging findings in the era of modern imaging modalities has enhanced our ability to detect cardiovascular involvement and lung metastases early and avoid misinterpreting tumor thrombus in draining systemic veins or pulmonary arteries as bland thrombus.

MR Imaging of Chest Wall Tumors.

Tumors of the chest wall are uncommon lesions that represent approximately 5% of all thoracic malignancies. These tumors comprise a heterogeneous group of neoplasms that may arise from osseous structures or soft tissues and may be malignant or benign. Most chest wall neoplasms are malignant and include lesions that secondarily involve the chest wall via direct invasion or metastasis from intrathoracic tumors or arise as primary tumors. More than 20% of lesions may be detected on chest radiography. This review focuses on key features of malignant and benign chest wall tumors (primary and secondary) on MR imaging examinations.