Artificial intelligence for prediction of COVID-19 progression using CT imaging and clinical data.

OBJECTIVES: Early recognition of coronavirus disease 2019 (COVID-19) severity can guide patient management. However, it is challenging to predict when COVID-19 patients will progress to critical illness. This study aimed to develop an artificial intelligence system to predict future deterioration to critical illness in COVID-19 patients. METHODS: An artificial intelligence (AI) system in a time-to-event analysis framework was developed to integrate chest CT and clinical data for risk prediction of future deterioration to critical illness in patients with COVID-19. RESULTS: A multi-institutional international cohort of 1,051 patients with RT-PCR confirmed COVID-19 and chest CT was included in this study. Of them, 282 patients developed critical illness, which was defined as requiring ICU admission and/or mechanical ventilation and/or reaching death during their hospital stay. The AI system achieved a C-index of 0.80 for predicting individual COVID-19 patients' to critical illness. The AI system successfully stratified the patients into high-risk and low-risk groups with distinct progression risks (p < 0.0001). CONCLUSIONS: Using CT imaging and clinical data, the AI system successfully predicted time to critical illness for individual patients and identified patients with high risk. AI has the potential to accurately triage patients and facilitate personalized treatment. KEY POINT: • AI system can predict time to critical illness for patients with COVID-19 by using CT imaging and clinical data.

Neoadjuvant radiation for cutaneous and soft tissue angiosarcoma.

BACKGROUND AND OBJECTIVES: Neoadjuvant radiation (NRT) is frequently utilized in soft tissue sarcomas to increase local control. Its utility in cutaneous and soft tissue angiosarcoma remains poorly defined. METHODS: This retrospective cohort study was performed using the National Cancer Database (2004-2016) evaluating patients with clinically localized, surgically resected angiosarcomas. Factors associated with receipt of NRT in the overall cohort and margin positivity in treatment naïve patients were identified by univariate and multivariable logistic regression analyses. Survival was assessed using Kaplan-Meier analysis. RESULTS: Of 597 patients, 27 (4.5%) received NRT. Increasing age (odds ratio [OR] 0.95, p = 0.025), tumor size more than or equal to 5 cm (OR 3.16, p = 0.02), and extremity tumor location (OR 3.99, p = 0.04) were associated with receipt of NRT. All patients who received NRT achieved an R0 resection (p = 0.03) compared with 17.9% of patients without NRT. Factors associated with risk of margin positivity included tumor size more than or equal to 5 cm (OR 1.85, p = 0.01), and head/neck location (OR 2.24, p = 0.006). NRT was not significantly associated with improved survival (p = 0.21). CONCLUSIONS: NRT improves rates of R0 resection but is infrequently utilized in cutaneous and soft tissue angiosarcoma. Increased usage of NRT, particularly for patients with lesions more than or equal to 5 cm, or head and neck location, may help achieve complete resections.

Prognostication of patients with COVID-19 using artificial intelligence based on chest x-rays and clinical data: a retrospective study.

BACKGROUND: Chest x-ray is a relatively accessible, inexpensive, fast imaging modality that might be valuable in the prognostication of patients with COVID-19. We aimed to develop and evaluate an artificial intelligence system using chest x-rays and clinical data to predict disease severity and progression in patients with COVID-19. METHODS: We did a retrospective study in multiple hospitals in the University of Pennsylvania Health System in Philadelphia, PA, USA, and Brown University affiliated hospitals in Providence, RI, USA. Patients who presented to a hospital in the University of Pennsylvania Health System via the emergency department, with a diagnosis of COVID-19 confirmed by RT-PCR and with an available chest x-ray from their initial presentation or admission, were retrospectively identified and randomly divided into training, validation, and test sets (7:1:2). Using the chest x-rays as input to an EfficientNet deep neural network and clinical data, models were trained to predict the binary outcome of disease severity (ie, critical or non-critical). The deep-learning features extracted from the model and clinical data were used to build time-to-event models to predict the risk of disease progression. The models were externally tested on patients who presented to an independent multicentre institution, Brown University affiliated hospitals, and compared with severity scores provided by radiologists. FINDINGS: 1834 patients who presented via the University of Pennsylvania Health System between March 9 and July 20, 2020, were identified and assigned to the model training (n=1285), validation (n=183), or testing (n=366) sets. 475 patients who presented via the Brown University affiliated hospitals between March 1 and July 18, 2020, were identified for external testing of the models. When chest x-rays were added to clinical data for severity prediction, area under the receiver operating characteristic curve (ROC-AUC) increased from 0·821 (95% CI 0·796-0·828) to 0·846 (0·815-0·852; p<0·0001) on internal testing and 0·731 (0·712-0·738) to 0·792 (0·780-0 ·803; p<0·0001) on external testing. When deep-learning features were added to clinical data for progression prediction, the concordance index (C-index) increased from 0·769 (0·755-0·786) to 0·805 (0·800-0·820; p<0·0001) on internal testing and 0·707 (0·695-0·729) to 0·752 (0·739-0·764; p<0·0001) on external testing. The image and clinical data combined model had significantly better prognostic performance than combined severity scores and clinical data on internal testing (C-index 0·805 vs 0·781; p=0·0002) and external testing (C-index 0·752 vs 0·715; p<0·0001). INTERPRETATION: In patients with COVID-19, artificial intelligence based on chest x-rays had better prognostic performance than clinical data or radiologist-derived severity scores. Using artificial intelligence, chest x-rays can augment clinical data in predicting the risk of progression to critical illness in patients with COVID-19. FUNDING: Brown University, Amazon Web Services Diagnostic Development Initiative, Radiological Society of North America, National Cancer Institute and National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health.

Preoperative prediction of the stage, size, grade, and necrosis score in clear cell renal cell carcinoma using MRI-based radiomics.

PURPOSE: Clear cell renal cell carcinoma (ccRCC) is the most common subtype of renal cell carcinoma. Currently, there is a lack of noninvasive methods to stratify ccRCC prognosis prior to any invasive therapies. The purpose of this study was to preoperatively predict the tumor stage, size, grade, and necrosis (SSIGN) score of ccRCC using MRI-based radiomics. METHODS: A multicenter cohort of 364 histopathologically confirmed ccRCC patients (272 low [< 4] and 92 high [≥ 4] SSIGN score) with preoperative T2-weighted and T1-contrast-enhanced MRI were retrospectively identified and divided into training (254 patients) and testing sets (110 patients). The performance of a manually optimized radiomics model was assessed by measuring accuracy, sensitivity, specificity, area under receiver operating characteristic curve (AUROC), and area under precision-recall curve (AUPRC) on an independent test set, which was not included in model training. Lastly, its performance was compared to that of a machine learning pipeline, Tree-Based Pipeline Optimization Tool (TPOT). RESULTS: The manually optimized radiomics model using Random Forest classification and Analysis of Variance feature selection methods achieved an AUROC of 0.89, AUPRC of 0.81, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. The TPOT using Extra Trees Classifier achieved an AUROC of 0.94, AUPRC of 0.83, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. CONCLUSION: Preoperative MR radiomics can accurately predict SSIGN score of ccRCC, suggesting its promise as a prognostic tool that can be used in conjunction with diagnostic markers.

Gene Fusions in Ocular Adnexal Tumors.

PURPOSE: To highlight the increasing importance of gene fusions in the diagnosis, prognosis, and therapy of ocular adnexal tumors. DESIGN: Perspective. METHODS: A focused review of gene fusions, their pathogenic mechanism, and gene fusion detection methods in lacrimal gland and primary orbital and ocular adnexal soft tissue tumors; reappraisal of diagnostic, prognostic, and therapeutic approach to ocular adnexal tumors in light of emerging molecular genetic data. RESULTS: The widespread implementation of fluorescence in situ hybridization and next-generation sequencing methods in pathology practice has led to identification of recurrent gene rearrangements and fusions in a variety of tumors. As a result, molecular genetic methods have become the gold standard for diagnosis of tumors with overlapping histology and immunophenotype, such as small round blue cell tumors. Identification of canonic gene fusions has led to development of sensitive and specific immunohistochemical markers, such as STAT6 in solitary fibrous tumor. In addition to diagnostic accuracy, gene fusions have prognostic implications, such as unfavorable prognosis of PAX3-FOXO1 fusion in alveolar rhabdomyosarcoma. Finally, recognition of gene fusions as a driving mechanism in neoplasia has led to development of U.S. Food and Drug Administration-approved targeted therapies, such as TRK inhibitors for NTRK fusion-positive cancers. CONCLUSION: The discovery of recurrent gene fusions in various tumors, including those involving ocular adnexa, has led to a deeper insight into the molecular mechanisms of these neoplasms, revolutionizing our approach to their diagnosis, prognostication, and therapy.

Primary Pulmonary Myxoid Sarcoma and Myxoid Angiomatoid Fibrous Histiocytoma: A Unifying Continuum With Shared and Distinct Features.

Primary pulmonary myxoid sarcoma (PPMS) is a recently reported, exceedingly rare low-grade lung neoplasm characterized by reticular/lace-like growth of spindle to epithelioid cells embedded in an abundant myxoid matrix. Morphologically, it overlaps with a myxoid variant of angiomatoid fibrous histiocytoma (AFH) of the soft tissue. Genetically, they were both reported to harbor EWSR1-CREB1 fusion, while EWSR1-ATF1 has only been reported in AFH thus far. We report a case of primary pulmonary low-grade myxoid spindle cell tumor with morphologic and immunohistochemical features of PPMS but with an EWSR1-ATF1 fusion gene. In addition, we also encountered a case of endobronchial AFH with EWSR1-CREB1 translocation but also focal morphologic features of PPMS. These findings provide new evidence supporting the concept that PPMS and a myxoid variant of AFH represent a continuum with overlapping histologic, immunohistochemical, and genetic features.

Atypical Lipomatous Tumor/Well-Differentiated Liposarcoma with Myxoid Stroma in a Hereditary Retinoblastoma Survivor.

Atypical lipomatous tumor/well-differentiated liposarcoma (ALT/WDL) is an indolent, locally aggressive mesenchymal neoplasm, most often confined to the lower extremities and retroperitoneum and rarely identified in the orbit. Diagnosis of ALT/WDL can be challenging due to its frequent morphologic overlap with benign adipose lesions and other more aggressive liposarcoma subtypes, including myxoid liposarcoma. We describe a 26-year-old female with a history of hereditary retinoblastoma and external-beam radiotherapy to the orbit, who developed orbital liposarcoma. Although initial morphologic assessment raised the consideration of myxoid liposarcoma, subsequent fluorescein in situ hybridization studies demonstrated MDM2 and DDIT3 coamplification without DDIT3 rearrangement, supporting the diagnosis of ALT/WDL with myxoid stroma. The literature review of previously reported orbital myxoid liposarcomas revealed a morphologic overlap of documented tumors with ALT/WDL, dedifferentiated liposarcoma, and pleomorphic liposarcoma with myxoid stroma as well as an absence of immunohistochemical and molecular genetic data supportive of the diagnosis of myxoid liposarcoma. This case emphasizes the potential overlap of ALT/WDL with myxoid liposarcoma and the increasing importance of molecular genetic studies in the diagnosis, prognosis, and management of orbital liposarcoma.

Mucoepidermoid Carcinoma of the Lacrimal Sac: Clinical-Pathologic Analysis, Including Molecular Genetics.

PURPOSE: The aim of this study was to assess whether mucoepidermoid carcinoma of the lacrimal sac is a counterpart of CRTC1/3-MAML2 gene fusion-related salivary gland mucoepidermoid carcinoma. METHODS: In this retrospective observational case series, pathology records were searched for all cases of lacrimal sac mucoepidermoid carcinoma diagnosed between 1990 and 2018. Data collected included demographics, clinical findings, management, and follow-up. Pathologic parameters assessed included tumor morphology, immunohistochemistry, and MAML2 and EGFR fluorescence in situ hybridization (FISH) studies. RESULTS: Six patients with mucoepidermoid carcinoma of the lacrimal sac, 5 males and 1 female, with a median age of 63 years (range 24-66) were identified. Five tumors were managed with radical resection and 1 patient underwent orbital exenteration. None of the patients developed recurrence or metastases with an average follow-up of 18 months (range 13-23). All tumors had morphologic and immunohistochemical features of mucoepidermoid carcinoma and overexpressed EGFR. MAML2 FISH was negative for MAML2 rearrangement in all tumors. EGFR FISH demonstrated EGFR amplification in 1 tumor. CONCLUSIONS: Mucoepidermoid carcinoma of the lacrimal sac is not a lacrimal sac counterpart of CRTC1/3-MAML2 gene fusion-related salivary gland mucoepidermoid carcinoma. EGFR pathway activation and EGFR amplification in a subset of these neoplasms suggest the potential role for anti-EGFR agents.

Conjunctival 'mucoepidermoid carcinoma' revisited: a revision of terminology, based on morphologic, immunohistochemical and molecular findings of 14 cases, and the 2018 WHO Classification of Tumours of the Eye.

In 2018, the consensus meeting for the WHO Classification of Tumours of the Eye decided that conjunctival mucoepidermoid carcinoma should be reclassified as adenosquamous carcinoma, as this represented a better morphological fit. To examine the applicability of this terminology, we studied the clinical, histopathological, immunohistochemical and molecular pathology of 14 cases that were originally diagnosed as conjunctival mucoepidermoid carcinoma. There were 7 (50%) females and 7 (50%) males. The median age was 64 years. The left eye was affected in 8 and the right eye in 6 patients. In-situ carcinoma was present in 11/14 (79%) cases and comprised in-situ squamous cell carcinoma (SCC) and conjunctival intraepithelial neoplasia with mucinous differentiation (CIN-Muc). Invasive carcinoma was present in 11/14 (79%) cases. Group 1 (1/11 cases, 9%) comprised invasive SCC only. Group 2 (6/11 cases, 55%) comprised SCC with mucinous differentiation, manifesting as scattered intracellular mucin, occasionally together with intercellular mucin, with no evidence of true glandular differentiation. Group 3 (3/11 cases. 27%) comprised true adenosquamous carcinoma. Group 4 (1/11 cases, 9%) comprised pure adenocarcinoma. Thirteen of 14 cases (93%) underwent FISH for MAML2 translocation and none were rearranged. Two cases harboured high-risk HPV (type 16 and 18). The combined findings confirm that all lesions in our study were not mucoepidermoid carcinoma, but represented predominantly SCC with mucinous differentiation and adenosquamous carcinoma. We, therefore, recommend future revision of the WHO classification to include SCC with mucinous differentiation alongside adenosquamous carcinoma.

Tubular apocrine adenoma of the eyelid - A case report and literature review.

Tubular apocrine adenoma is a rare benign adnexal neoplasm most commonly identified in the scalp, composed of a dermal proliferation of apocrine tubules in a background of hyalinized stroma. Tubular apocrine adenoma can be a component of various sweat gland tumors and can also morphologically overlap with other sweat gland neoplasms. Isolated tubular apocrine adenoma arising in the glands of Moll is exceedingly rare, with only 4 previously reported cases. We present a 63-year-old male with tubular apocrine adenoma of the left upper eyelid, which recurred following initial incomplete excision. Although the lesion showed focal morphologic similarity to the apocrine variant of pleomorphic adenoma (chondroid syringoma), the diagnosis of tubular apocrine adenoma was supported by fluorescence in situ hybridization studies, which demonstrated absence of PLAG1 and HMGA2 gene rearrangements seen in pleomorphic adenoma. This case illustrates the clinical, microscopic and immunohistochemical features of tubular apocrine adenoma. The recent advances in our understanding of the molecular genetics of tubular apocrine adenoma and related tumors, and how these advances shape the evolving classification of sweat gland tumors are reviewed.

Orbital Atypical Lipomatous Tumor/Well-Differentiated Liposarcoma Masquerading as Pleomorphic Lipoma: A Diagnostic Challenge.

Atypical lipomatous tumor/well-differentiated liposarcoma is a common neoplasm of the superficial and deep soft tissues of the extremities, trunk, and retroperitoneum. Atypical lipomatous tumor/well-differentiated liposarcoma is very rare in the orbit, with only 19 previously reported cases. The authors describe a 22-year-old woman who presented with an 8-month history of diplopia and was found to have an orbital mass on MRI. The excised tumor initially was interpreted as spindle cell/pleomorphic lipoma based on its morphologic and immunohistochemical features. Nine years later, the patient returned with a recurrence that required surgical debulking. Histopathologic and molecular cytogenetic evaluation of both primary and recurrent lesions disclosed Atypical lipomatous tumor/well-differentiated liposarcoma. This case highlights the diagnostic challenges and the importance of molecular genetic studies in evaluation of fatty orbital tumors.

Combined Sclerosing and Spindle Cell Rhabdomyosarcoma in Previous Craniotomy Site: A Case Report and a Review of the Literature.

Sclerosing rhabdomyosarcoma (RMS) is a rare subtype of RMS with unique prominent stromal hyalinization and a pseudovascular architecture. It overlaps morphologically with spindle cell RMS and poses both diagnostic and therapeutic challenges because of its rarity and aggressive clinical course. In this article, we report a case of sclerosing RMS arising from a prior craniotomy site, which demonstrated both sclerosing and spindle cell components. A literature review of RMS with sclerosing morphology identified 122 cases. Our review documents the following: sclerosing RMS occurs in both childhood and adult populations, has a predilection for the head and neck areas, and has a worse prognosis in adults. Sclerosing RMS harbors a high frequency of MYOD1 mutations, conferring a poor clinical outcome. Sclerosing RMS and spindle RMS likely represent a morphologic spectrum of one entity.

Adenocarcinoma, Not Otherwise Specified, Arising in Accessory Lacrimal Gland: A Diagnostic Challenge.

Neoplasms arising in accessory lacrimal glands are rare. We describe a 33-year-old man with adenocarcinoma arising in the left lower eyelid accessory lacrimal gland. Microscopic evaluation demonstrated an infiltrative neoplasm composed of mildly to moderately pleomorphic cells with abundant eosinophilic cytoplasm and focal intracytoplasmic vacuoles, arranged predominantly in ductules. Foci of luminal and intracytoplasmic eosinophilic secretory material and occasional mucin were noted. An in situ component was identified in the gland of Wolfring. Though perineural invasion was present, high-grade nuclear features, brisk mitotic activity, and comedonecrosis were not identified. Immunohistochemical studies were notable for immunoreactivity of the tumor cells for CK7, carcinoembryonic antigen, BRST-2, androgen receptors, and HER2/neu (2+). The neoplastic cells were negative for CK20, estrogen and progesterone receptors, S-100, p63, calponin, thyroid transcription factor-1, and prostate-specific antigen. Fluorescence in situ hybridization studies for ETV6 and MAML2 rearrangements and for HER2/neu amplification were negative. Because of the absence of unifying morphologic, immunophenotypic, and molecular genetic findings, the diagnosis of adenocarcinoma, not otherwise specified, was rendered. The patient underwent comprehensive oncologic workup, which was negative for another primary tumor and metastases. He remains disease free with a follow-up of 4 years. This case illustrates the challenges encountered in applying salivary gland tumor classification to the accessory lacrimal gland neoplasm.

Secretory Carcinoma of the Lacrimal Gland: A Rare Case Report.

Secretory carcinoma is a salivary gland malignancy that recapitulates secretory carcinoma of the breast, along with its shared ETV6-NTRK3 gene fusion. Characterization of histopathologic, immunohistochemical, and molecular genetic features of this neoplasm has led to reclassification of a heterogeneous group of salivary gland carcinomas as secretory carcinoma and to identification of this neoplasm in other gland-containing tissues. The authors describe a 52-year-old man who presented with a 2-week history of diplopia and a well-circumscribed right orbital mass. The tumor was resected via lateral orbitotomy approach. Pathologic evaluation demonstrated secretory carcinoma, previously not described in the main lacrimal gland. Recognition of lacrimal gland secretory carcinoma may lead to reappraisal of morphologically similar, but biologically heterogeneous lacrimal gland neoplasms, providing an insight into this tumor's clinical presentation and prognosis. Accurate diagnosis of this malignancy has important management and prognostic implications, particularly with emergence of targeted therapies.

ATF4-dependent induction of heme oxygenase 1 prevents anoikis and promotes metastasis.

The integrated stress response (ISR) is a critical mediator of cancer cell survival, and targeting the ISR inhibits tumor progression. Here, we have shown that activating transcription factor 4 (ATF4), a master transcriptional effector of the ISR, protects transformed cells against anoikis - a specialized form of apoptosis - following matrix detachment and also contributes to tumor metastatic properties. Upon loss of attachment, ATF4 activated a coordinated program of cytoprotective autophagy and antioxidant responses, including induced expression of the major antioxidant enzyme heme oxygenase 1 (HO-1). HO-1 upregulation was the result of simultaneous activation of ATF4 and the transcription factor NRF2, which converged on the HO1 promoter. Increased levels of HO-1 ameliorated oxidative stress and cell death. ATF4-deficient human fibrosarcoma cells were unable to colonize the lungs in a murine model, and reconstitution of ATF4 or HO-1 expression in ATF4-deficient cells blocked anoikis and rescued tumor lung colonization. HO-1 expression was higher in human primary and metastatic tumors compared with noncancerous tissue. Moreover, HO-1 expression correlated with reduced overall survival of patients with lung adenocarcinoma and glioblastoma. These results establish HO-1 as a mediator of ATF4-dependent anoikis resistance and tumor metastasis and suggest ATF4 and HO-1 as potential targets for therapeutic intervention in solid tumors.

Hemangiomas of the brachial plexus: a case series.

OBJECTIVE: Hemangiomas of the brachial plexus are very rare, and there has not been a collection of multiple cases published in the literature to date. Extraneural brachial plexus hemangiomas typically present with similar signs and symptoms as nerve sheath tumors, including pain, paresthesia, and occasionally weakness, in addition to nonspecific imaging findings, making their diagnosis difficult. Exploratory surgery can lead to significant bleeding and nerve injury when a hemangioma or an associated aneurysm is encountered intraoperatively. We present 5 cases of extraneural hemangiomas causing brachial plexopathy, including pre-, intra-, and postoperative decision making, with an emphasis on diagnostic and management issues as well as outcomes. METHODS: A retrospective review was performed of 5 patients who underwent surgery at a university teaching hospital between 1995 and 2007 for exploration of brachial plexus lesions that were confirmed to be hemangiomas at pathological examination. RESULTS: All 5 patients presented with findings on history, physical examination, imaging, and electromyography suggesting a diagnosis of nerve sheath tumor. Two patients had biopsies (1 needle, 1 open), both of which were nondiagnostic. Three patients underwent digital subtraction angiography with successful preoperative embolization. Each patient had a complete or a radical subtotal tumor resection, and all were intact neurologically after surgical resection. Pathological evaluation identified 3 venous hemangiomas, 1 hemangioma with arteriovenous malformation features, and 1 Masson hemangioma associated with a large aneurysm. CONCLUSION: Extraneural hemangiomas of the brachial plexus are very rare, but a high index of suspicion and appropriate preoperative evaluation, including angiography with the option for embolization, can result in decreased intraoperative hemorrhage and better patient outcomes.