Pathomic Features Reveal Immune and Molecular Evolution From Lung Preneoplasia to Invasive Adenocarcinoma.

Recent statistics on lung cancer, including the steady decline of advanced diseases and the dramatically increasing detection of early-stage diseases and indeterminate pulmonary nodules, mark the significance of a comprehensive understanding of early lung carcinogenesis. Lung adenocarcinoma (ADC) is the most common histologic subtype of lung cancer, and atypical adenomatous hyperplasia is the only recognized preneoplasia to ADC, which may progress to adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA) and eventually to invasive ADC. Although molecular evolution during early lung carcinogenesis has been explored in recent years, the progress has been significantly hindered, largely due to insufficient materials from ADC precursors. Here, we employed state-of-the-art deep learning and artificial intelligence techniques to robustly segment and recognize cells on routinely used hematoxylin and eosin histopathology images and extracted 9 biology-relevant pathomic features to decode lung preneoplasia evolution. We analyzed 3 distinct cohorts (Japan, China, and United States) covering 98 patients, 162 slides, and 669 regions of interest, including 143 normal, 129 atypical adenomatous hyperplasia, 94 AIS, 98 MIA, and 205 ADC. Extracted pathomic features revealed progressive increase of atypical epithelial cells and progressive decrease of lymphocytic cells from normal to AAH, AIS, MIA, and ADC, consistent with the results from tissue-consuming and expensive molecular/immune profiling. Furthermore, pathomics analysis manifested progressively increasing cellular intratumor heterogeneity along with the evolution from normal lung to invasive ADC. These findings demonstrated the feasibility and substantial potential of pathomics in studying lung cancer carcinogenesis directly from the low-cost routine hematoxylin and eosin staining.

Effects of prenatal exposure to maternal COVID-19 and perinatal care on neonatal outcome: results from the INTERCOVID Multinational Cohort Study.

BACKGROUND: The effect of COVID-19 in pregnancy on maternal outcomes and its association with preeclampsia and gestational diabetes mellitus have been reported; however, a detailed understanding of the effects of maternal positivity, delivery mode, and perinatal practices on fetal and neonatal outcomes is urgently needed. OBJECTIVE: To evaluate the impact of COVID-19 on fetal and neonatal outcomes and the role of mode of delivery, breastfeeding, and early neonatal care practices on the risk of mother-to-child transmission. STUDY DESIGN: In this cohort study that took place from March 2020 to March 2021, involving 43 institutions in 18 countries, 2 unmatched, consecutive, unexposed women were concomitantly enrolled immediately after each infected woman was identified, at any stage of pregnancy or delivery, and at the same level of care to minimize bias. Women and neonates were followed up until hospital discharge. COVID-19 in pregnancy was determined by laboratory confirmation and/or radiological pulmonary findings or ≥2 predefined COVID-19 symptoms. The outcome measures were indices of neonatal and perinatal morbidity and mortality, neonatal positivity and its correlation with mode of delivery, breastfeeding, and hospital neonatal care practices. RESULTS: A total of 586 neonates born to women with COVID-19 diagnosis and 1535 neonates born to women without COVID-19 diagnosis were enrolled. Women with COVID-19 diagnosis had a higher rate of cesarean delivery (52.8% vs 38.5% for those without COVID-19 diagnosis, P<.01) and pregnancy-related complications, such as hypertensive disorders of pregnancy and fetal distress (all with P<.001), than women without COVID-19 diagnosis. Maternal diagnosis of COVID-19 carried an increased rate of preterm birth (P≤.001) and lower neonatal weight (P≤.001), length, and head circumference at birth. In mothers with COVID-19 diagnosis, the length of in utero exposure was significantly correlated to the risk of the neonate testing positive (odds ratio, 4.5; 95% confidence interval, 2.2-9.4 for length of in utero exposure >14 days). Among neonates born to mothers with COVID-19 diagnosis, birth via cesarean delivery was a risk factor for testing positive for COVID-19 (odds ratio, 2.4; 95% confidence interval, 1.2-4.7), even when severity of maternal conditions was considered and after multivariable logistic analysis. In the subgroup of neonates born to women with COVID-19 diagnosis, the outcomes worsened when the neonate also tested positive, with higher rates of neonatal intensive care unit admission, fever, gastrointestinal and respiratory symptoms, and death, even after adjusting for prematurity. Breastfeeding by mothers with COVID-19 diagnosis and hospital neonatal care practices, including immediate skin-to-skin contact and rooming-in, were not associated with an increased risk of newborn positivity. CONCLUSION: In this multinational cohort study, COVID-19 in pregnancy was associated with increased maternal and neonatal complications. Cesarean delivery was significantly associated with newborn COVID-19 diagnosis. Vaginal delivery should be considered the safest mode of delivery if obstetrical and health conditions allow it. Mother-to-child skin-to-skin contact, rooming-in, and direct breastfeeding were not risk factors for newborn COVID-19 diagnosis, thus well-established best practices can be continued among women with COVID-19 diagnosis.

Diabetes mellitus, maternal adiposity, and insulin-dependent gestational diabetes are associated with COVID-19 in pregnancy: the INTERCOVID study.

BACKGROUND: Among nonpregnant individuals, diabetes mellitus and high body mass index increase the risk of COVID-19 and its severity. OBJECTIVE: This study aimed to determine whether diabetes mellitus and high body mass index are risk factors for COVID-19 in pregnancy and whether gestational diabetes mellitus is associated with COVID-19 diagnosis. STUDY DESIGN: INTERCOVID was a multinational study conducted between March 2020 and February 2021 in 43 institutions from 18 countries, enrolling 2184 pregnant women aged ≥18 years; a total of 2071 women were included in the analyses. For each woman diagnosed with COVID-19, 2 nondiagnosed women delivering or initiating antenatal care at the same institution were also enrolled. The main exposures were preexisting diabetes mellitus, high body mass index (overweight or obesity was defined as a body mass index ≥25 kg/m2), and gestational diabetes mellitus in pregnancy. The main outcome was a confirmed diagnosis of COVID-19 based on a real-time polymerase chain reaction test, antigen test, antibody test, radiological pulmonary findings, or ≥2 predefined COVID-19 symptoms at any time during pregnancy or delivery. Relationships of exposures and COVID-19 diagnosis were assessed using generalized linear models with a Poisson distribution and log link function, with robust standard errors to account for model misspecification. Furthermore, we conducted sensitivity analyses: (1) restricted to those with a real-time polymerase chain reaction test or an antigen test in the last week of pregnancy, (2) restricted to those with a real-time polymerase chain reaction test or an antigen test during the entire pregnancy, (3) generating values for missing data using multiple imputation, and (4) analyses controlling for month of enrollment. In addition, among women who were diagnosed with COVID-19, we examined whether having gestational diabetes mellitus, diabetes mellitus, or high body mass index increased the risk of having symptomatic vs asymptomatic COVID-19. RESULTS: COVID-19 was associated with preexisting diabetes mellitus (risk ratio, 1.94; 95% confidence interval, 1.55-2.42), overweight or obesity (risk ratio, 1.20; 95% confidence interval, 1.06-1.37), and gestational diabetes mellitus (risk ratio, 1.21; 95% confidence interval, 0.99-1.46). The gestational diabetes mellitus association was specifically among women requiring insulin, whether they were of normal weight (risk ratio, 1.79; 95% confidence interval, 1.06-3.01) or overweight or obese (risk ratio, 1.77; 95% confidence interval, 1.28-2.45). A somewhat stronger association with COVID-19 diagnosis was observed among women with preexisting diabetes mellitus, whether they were of normal weight (risk ratio, 1.93; 95% confidence interval, 1.18-3.17) or overweight or obese (risk ratio, 2.32; 95% confidence interval, 1.82-2.97). When the sample was restricted to those with a real-time polymerase chain reaction test or an antigen test in the week before delivery or during the entire pregnancy, including missing variables using imputation or controlling for month of enrollment, the observed associations were comparable. CONCLUSION: Diabetes mellitus and overweight or obesity were risk factors for COVID-19 diagnosis in pregnancy, and insulin-dependent gestational diabetes mellitus was associated with the disease. Therefore, it is essential that women with these comorbidities are vaccinated.

Preeclampsia and COVID-19: results from the INTERCOVID prospective longitudinal study.

BACKGROUND: It is unclear whether the suggested link between COVID-19 during pregnancy and preeclampsia is an independent association or if these are caused by common risk factors. OBJECTIVE: This study aimed to quantify any independent association between COVID-19 during pregnancy and preeclampsia and to determine the effect of these variables on maternal and neonatal morbidity and mortality. STUDY DESIGN: This was a large, longitudinal, prospective, unmatched diagnosed and not-diagnosed observational study assessing the effect of COVID-19 during pregnancy on mothers and neonates. Two consecutive not-diagnosed women were concomitantly enrolled immediately after each diagnosed woman was identified, at any stage during pregnancy or delivery, and at the same level of care to minimize bias. Women and neonates were followed until hospital discharge using the standardized INTERGROWTH-21st protocols and electronic data management system. A total of 43 institutions in 18 countries contributed to the study sample. The independent association between the 2 entities was quantified with the risk factors known to be associated with preeclampsia analyzed in each group. The outcomes were compared among women with COVID-19 alone, preeclampsia alone, both conditions, and those without either of the 2 conditions. RESULTS: We enrolled 2184 pregnant women; of these, 725 (33.2%) were enrolled in the COVID-19 diagnosed and 1459 (66.8%) in the COVID-19 not-diagnosed groups. Of these women, 123 had preeclampsia of which 59 of 725 (8.1%) were in the COVID-19 diagnosed group and 64 of 1459 (4.4%) were in the not-diagnosed group (risk ratio, 1.86; 95% confidence interval, 1.32-2.61). After adjustment for sociodemographic factors and conditions associated with both COVID-19 and preeclampsia, the risk ratio for preeclampsia remained significant among all women (risk ratio, 1.77; 95% confidence interval, 1.25-2.52) and nulliparous women specifically (risk ratio, 1.89; 95% confidence interval, 1.17-3.05). There was a trend but no statistical significance among parous women (risk ratio, 1.64; 95% confidence interval, 0.99-2.73). The risk ratio for preterm birth for all women diagnosed with COVID-19 and preeclampsia was 4.05 (95% confidence interval, 2.99-5.49) and 6.26 (95% confidence interval, 4.35-9.00) for nulliparous women. Compared with women with neither condition diagnosed, the composite adverse perinatal outcome showed a stepwise increase in the risk ratio for COVID-19 without preeclampsia, preeclampsia without COVID-19, and COVID-19 with preeclampsia (risk ratio, 2.16; 95% confidence interval, 1.63-2.86; risk ratio, 2.53; 95% confidence interval, 1.44-4.45; and risk ratio, 2.84; 95% confidence interval, 1.67-4.82, respectively). Similar findings were found for the composite adverse maternal outcome with risk ratios of 1.76 (95% confidence interval, 1.32-2.35), 2.07 (95% confidence interval, 1.20-3.57), and 2.77 (95% confidence interval, 1.66-4.63). The association between COVID-19 and gestational hypertension and the direction of the effects on preterm birth and adverse perinatal and maternal outcomes, were similar to preeclampsia, but confined to nulliparous women with lower risk ratios. CONCLUSION: COVID-19 during pregnancy is strongly associated with preeclampsia, especially among nulliparous women. This association is independent of any risk factors and preexisting conditions. COVID-19 severity does not seem to be a factor in this association. Both conditions are associated independently of and in an additive fashion with preterm birth, severe perinatal morbidity and mortality, and adverse maternal outcomes. Women with preeclampsia should be considered a particularly vulnerable group with regard to the risks posed by COVID-19.

Kidney care in low- and middle-income countries.

Optimal kidney care requires a trained nephrology workforce, essential healthcare services, and medications. This study aimed to identify the access to these resources on a global scale using data from the multinational survey conducted by the International Society of Nephrology (ISN) (Global Kidney Health Atlas (GKHA) project), with emphasis on developing nations. For data analysis, the 125 participating countries were sorted into the 4 World Bank income groups: low income (LIC), lower-middle income (LMIC), upper-middle income (UMIC), and high income (HIC). A severe shortage of nephrologists was observed in LIC and LMIC with < 5 nephrologists per million population. Many LIC were unable to access estimated glomerular filtration rate (eGFR) and albuminuria (proteinuria) tests in primary-care levels. Acute and chronic hemodialysis was available in most countries, although acute and chronic peritoneal dialysis access was severely limited in LIC (24% and 35%, respectively). Most countries had kidney transplantation access, except for LIC (12%). HIC and UMIC funded their renal replacement therapy (RRT) and renal medications primarily through public means, whereas LMIC and LIC required private and out-of-pocket contributions. In conclusion, this study found a huge gap in the availability and access to trained nephrology workforce, tools for diagnosis and management of CKD, RRT, and funding of RRT and essential medications in LIC and LMIC.

Tubal ectopic gestation associated with genital schistosomiasis: a case report.

Schistosoma are trematode blood flukes of the family Schistosomidae affecting the urinary and gastro-intestinal tracts. Riverine areas of the world such as in Africa, Eastern Mediterranean, Central American and East Asia are endemic for the disease, with S. haematobium accounting for most of the symptomatic genital infection. A case of a 25-year-old woman with 8 weeks amenorrhoea, lower abdominal pain and per vaginal bleeding was managed for ruptured ectopic pregnancy and discovered to have tubal infection by Schistosoma on histological examination is presented.

CoCo-ST: Comparing and Contrasting Spatial Transcriptomics data sets using graph contrastive learning.

Traditional feature dimension reduction methods have been widely used to uncover biological patterns or structures within individual spatial transcriptomics data. However, these methods are designed to yield feature representations that emphasize patterns or structures with dominant high variance, such as the normal tissue spatial pattern in a precancer setting. Consequently, they may inadvertently overlook patterns of interest that are potentially masked by these high-variance structures. Herein we present our graph contrastive feature representation method called CoCo-ST (Comparing and Contrasting Spatial Transcriptomics) to overcome this limitation. By incorporating a background data set representing normal tissue, this approach enhances the identification of interesting patterns in a target data set representing precancerous tissue. Simultaneously, it mitigates the influence of dominant common patterns shared by the background and target data sets. This enables discerning biologically relevant features crucial for capturing tissue-specific patterns, a capability we showcased through the analysis of serial mouse precancerous lung tissue samples.

Joint multi-omics discriminant analysis with consistent representation learning using PANDA.

Integrative multi-omics analysis provides deeper insight and enables better and more realistic modeling of the underlying biology and causes of diseases than does single omics analysis. Although several integrative multi-omics analysis methods have been proposed and demonstrated promising results in integrating distinct omics datasets, inconsistent distribution of the different omics data, which is caused by technology variations, poses a challenge for paired integrative multi-omics methods. In addition, the existing discriminant analysis-based integrative methods do not effectively exploit correlation and consistent discriminant structures, necessitating a compromise between correlation and discrimination in using these methods. Herein we present PAN-omics Discriminant Analysis (PANDA), a joint discriminant analysis method that seeks omics-specific discriminant common spaces by jointly learning consistent discriminant latent representations for each omics. PANDA jointly maximizes between-class and minimizes within-class omics variations in a common space and simultaneously models the relationships among omics at the consistency representation and cross-omics correlation levels, overcoming the need for compromise between discrimination and correlation as with the existing integrative multi-omics methods. Because of the consistency representation learning incorporated into the objective function of PANDA, this method seeks a common discriminant space to minimize the differences in distributions among omics, can lead to a more robust latent representations than other methods, and is against the inconsistency of the different omics. We compared PANDA to 10 other state-of-the-art multi-omics data integration methods using both simulated and real-world multi-omics datasets and found that PANDA consistently outperformed them while providing meaningful discriminant latent representations. PANDA is implemented using both R and MATLAB, with codes available at https://github.com/WuLabMDA/PANDA.

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.

Enhancing NSCLC recurrence prediction with PET/CT habitat imaging, ctDNA, and integrative radiogenomics-blood insights.

While we recognize the prognostic importance of clinicopathological measures and circulating tumor DNA (ctDNA), the independent contribution of quantitative image markers to prognosis in non-small cell lung cancer (NSCLC) remains underexplored. In our multi-institutional study of 394 NSCLC patients, we utilize pre-treatment computed tomography (CT) and 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) to establish a habitat imaging framework for assessing regional heterogeneity within individual tumors. This framework identifies three PET/CT subtypes, which maintain prognostic value after adjusting for clinicopathologic risk factors including tumor volume. Additionally, these subtypes complement ctDNA in predicting disease recurrence. Radiogenomics analysis unveil the molecular underpinnings of these imaging subtypes, highlighting downregulation in interferon alpha and gamma pathways in the high-risk subtype. In summary, our study demonstrates that these habitat imaging subtypes effectively stratify NSCLC patients based on their risk levels for disease recurrence after initial curative surgery or radiotherapy, providing valuable insights for personalized treatment approaches.

Capacity for the management of kidney failure in the International Society of Nephrology Africa region: report from the 2023 ISN Global Kidney Atlas (ISN-GKHA).

The burden of chronic kidney disease and associated risk of kidney failure are increasing in Africa. The management of people with chronic kidney disease is fraught with numerous challenges because of limitations in health systems and infrastructures for care delivery. From the third iteration of the International Society of Nephrology Global Kidney Health Atlas, we describe the status of kidney care in the ISN Africa region using the World Health Organization building blocks for health systems. We identified limited government health spending, which in turn led to increased out-of-pocket costs for people with kidney disease at the point of service delivery. The health care workforce across Africa was suboptimal and further challenged by the exodus of trained health care workers out of the continent. Medical products, technologies, and services for the management of people with nondialysis chronic kidney disease and for kidney replacement therapy were scarce due to limitations in health infrastructure, which was inequitably distributed. There were few kidney registries and advocacy groups championing kidney disease management in Africa compared with the rest of the world. Strategies for ensuring improved kidney care in Africa include focusing on chronic kidney disease prevention and early detection, improving the effectiveness of the available health care workforce (e.g., multidisciplinary teams, task substitution, and telemedicine), augmenting kidney care financing, providing quality, up-to-date health information data, and improving the accessibility, affordability, and delivery of quality treatment (kidney replacement therapy or conservative kidney management) for all people living with kidney failure.

Post-dated Breech Pregnancy in a Non-obviously Communicating Rudimentary Horn of a Bicornuate Uterus Requiring Hemi-hysterectomy.

Developmental anomalies of the Müllerian duct systems such as the bicornuate uterus are rare globally and hardly do term pregnancies occur in conjunction with these abnormalities. The occurrence of post-dated pregnancy is rarely associated with a bicornuate uterus. We present a 35-year-old un-booked multigravida with post-date pregnancy complicated by breech and intrauterine foetal death (IUFD) in a rudimentary uterine horn. She had caesarean delivery complicated by intractable postpartum haemorrhage (PPH). This together with the risks of poor uterine involution in the postpartum and obstetric outcome in the event that another pregnancy occurs in the same horn subsequently warranted a caesarean hemi-hysterectomy of the rudimentary uterine horn. Uterine bicornuate is an uncommon genital tract anomaly and a rare cause of post-date pregnancy. Postpartum bleeding warranting caesarean hemi-hysterectomy should be anticipated as the pregnant horn may not be responsive to conventional oxytocics.

Phase engineering of nickel-based sulfides toward robust sodium-ion batteries.

Nickel-based sulfides are considered promising materials for sodium-ion batteries (SIBs) anodes due to their abundant resources and attractive theoretical capacity. However, their application is limited by slow diffusion kinetics and severe volume changes during cycling. Herein, we demonstrate a facile strategy for the synthesis of nitrogen-doped reduced graphene oxide (N-rGO) wrapped Ni3S2 nanocrystals composites (Ni3S2-N-rGO-700 °C) through the cubic NiS2 precursor under high temperature (700 ℃). Benefitting from the variation in crystal phase structure and robust coupling effect between the Ni3S2 nanocrystals and N-rGO matrix, the Ni3S2-N-rGO-700 °C exhibits enhanced conductivity, fast ion diffusion kinetics and outstanding structural stability. As a result, the Ni3S2-N-rGO-700 °C delivers excellent rate capability (345.17 mAh g-1 at a high current density of 5 A g-1) and long-term cyclic stability over 400 cycles at 2 A g-1 with a high reversible capacity of 377 mAh g-1 when evaluated as anodes for SIBs. This study open a promising avenue to realize advanced metal sulfide materials with desirable electrochemical activity and stability for energy storage applications.

Automated Cellular-Level Dual Global Fusion of Whole-Slide Imaging for Lung Adenocarcinoma Prognosis.

Histopathologic whole-slide images (WSI) are generally considered the gold standard for cancer diagnosis and prognosis. Survival prediction based on WSI has recently attracted substantial attention. Nevertheless, it remains a central challenge owing to the inherent difficulties of predicting patient prognosis and effectively extracting informative survival-specific representations from WSI with highly compounded gigapixels. In this study, we present a fully automated cellular-level dual global fusion pipeline for survival prediction. Specifically, the proposed method first describes the composition of different cell populations on WSI. Then, it generates dimension-reduced WSI-embedded maps, allowing for efficient investigation of the tumor microenvironment. In addition, we introduce a novel dual global fusion network to incorporate global and inter-patch features of cell distribution, which enables the sufficient fusion of different types and locations of cells. We further validate the proposed pipeline using The Cancer Genome Atlas lung adenocarcinoma dataset. Our model achieves a C-index of 0.675 (±0.05) in the five-fold cross-validation setting and surpasses comparable methods. Further, we extensively analyze embedded map features and survival probabilities. These experimental results manifest the potential of our proposed pipeline for applications using WSI in lung adenocarcinoma and other malignancies.

Heterogenous lung inflammation CT patterns distinguish pneumonia and immune checkpoint inhibitor pneumonitis and complement blood biomarkers in acute myeloid leukemia: proof of concept.

Background: Immune checkpoint inhibitors (ICI) may cause pneumonitis, resulting in potentially fatal lung inflammation. However, distinguishing pneumonitis from pneumonia is time-consuming and challenging. To fill this gap, we build an image-based tool, and further evaluate it clinically alongside relevant blood biomarkers. Materials and methods: We studied CT images from 97 patients with pneumonia and 29 patients with pneumonitis from acute myeloid leukemia treated with ICIs. We developed a CT-derived signature using a habitat imaging algorithm, whereby infected lungs are segregated into clusters ("habitats"). We validated the model and compared it with a clinical-blood model to determine whether imaging can add diagnostic value. Results: Habitat imaging revealed intrinsic lung inflammation patterns by identifying 5 distinct subregions, correlating to lung parenchyma, consolidation, heterogenous ground-glass opacity (GGO), and GGO-consolidation transition. Consequently, our proposed habitat model (accuracy of 79%, sensitivity of 48%, and specificity of 88%) outperformed the clinical-blood model (accuracy of 68%, sensitivity of 14%, and specificity of 85%) for classifying pneumonia versus pneumonitis. Integrating imaging and blood achieved the optimal performance (accuracy of 81%, sensitivity of 52% and specificity of 90%). Using this imaging-blood composite model, the post-test probability for detecting pneumonitis increased from 23% to 61%, significantly (p = 1.5E - 9) higher than the clinical and blood model (post-test probability of 22%). Conclusion: Habitat imaging represents a step forward in the image-based detection of pneumonia and pneumonitis, which can complement known blood biomarkers. Further work is needed to validate and fine tune this imaging-blood composite model and further improve its sensitivity to detect pneumonitis.

Efficacy and clinicogenomic correlates of response to immune checkpoint inhibitors alone or with chemotherapy in non-small cell lung cancer.

The role of combination chemotherapy with immune checkpoint inhibitors (ICI) (ICI-chemo) over ICI monotherapy (ICI-mono) in non-small cell lung cancer (NSCLC) remains underexplored. In this retrospective study of 1133 NSCLC patients, treatment with ICI-mono vs ICI-chemo associate with higher rates of early progression, but similar long-term progression-free and overall survival. Sequential vs concurrent ICI and chemotherapy have similar long-term survival, suggesting no synergism from combination therapy. Integrative modeling identified PD-L1, disease burden (Stage IVb; liver metastases), and STK11 and JAK2 alterations as features associate with a higher likelihood of early progression on ICI-mono. CDKN2A alterations associate with worse long-term outcomes in ICI-chemo patients. These results are validated in independent external (n = 89) and internal (n = 393) cohorts. This real-world study suggests that ICI-chemo may protect against early progression but does not influence overall survival, and nominates features that identify those patients at risk for early progression who may maximally benefit from ICI-chemo.

Predicting benefit from immune checkpoint inhibitors in patients with non-small-cell lung cancer by CT-based ensemble deep learning: a retrospective study.

BACKGROUND: Only around 20-30% of patients with non-small-cell lung cancer (NCSLC) have durable benefit from immune-checkpoint inhibitors. Although tissue-based biomarkers (eg, PD-L1) are limited by suboptimal performance, tissue availability, and tumour heterogeneity, radiographic images might holistically capture the underlying cancer biology. We aimed to investigate the application of deep learning on chest CT scans to derive an imaging signature of response to immune checkpoint inhibitors and evaluate its added value in the clinical context. METHODS: In this retrospective modelling study, 976 patients with metastatic, EGFR/ALK negative NSCLC treated with immune checkpoint inhibitors at MD Anderson and Stanford were enrolled from Jan 1, 2014, to Feb 29, 2020. We built and tested an ensemble deep learning model on pretreatment CTs (Deep-CT) to predict overall survival and progression-free survival after treatment with immune checkpoint inhibitors. We also evaluated the added predictive value of the Deep-CT model in the context of existing clinicopathological and radiological metrics. FINDINGS: Our Deep-CT model demonstrated robust stratification of patient survival of the MD Anderson testing set, which was validated in the external Stanford set. The performance of the Deep-CT model remained significant on subgroup analyses stratified by PD-L1, histology, age, sex, and race. In univariate analysis, Deep-CT outperformed the conventional risk factors, including histology, smoking status, and PD-L1 expression, and remained an independent predictor after multivariate adjustment. Integrating the Deep-CT model with conventional risk factors demonstrated significantly improved prediction performance, with overall survival C-index increases from 0·70 (clinical model) to 0·75 (composite model) during testing. On the other hand, the deep learning risk scores correlated with some radiomics features, but radiomics alone could not reach the performance level of deep learning, indicating that the deep learning model effectively captured additional imaging patterns beyond known radiomics features. INTERPRETATION: This proof-of-concept study shows that automated profiling of radiographic scans through deep learning can provide orthogonal information independent of existing clinicopathological biomarkers, bringing the goal of precision immunotherapy for patients with NSCLC closer. FUNDING: National Institutes of Health, Mark Foundation Damon Runyon Foundation Physician Scientist Award, MD Anderson Strategic Initiative Development Program, MD Anderson Lung Moon Shot Program, Andrea Mugnaini, and Edward L C Smith.

Chronic Lymphocytic Leukemia Progression Diagnosis with Intrinsic Cellular Patterns via Unsupervised Clustering.

Identifying the progression of chronic lymphocytic leukemia (CLL) to accelerated CLL (aCLL) or transformation to diffuse large B-cell lymphoma (Richter transformation; RT) has significant clinical implications as it prompts a major change in patient management. However, the differentiation between these disease phases may be challenging in routine practice. Unsupervised learning has gained increased attention because of its substantial potential in data intrinsic pattern discovery. Here, we demonstrate that cellular feature engineering, identifying cellular phenotypes via unsupervised clustering, provides the most robust analytic performance in analyzing digitized pathology slides (accuracy = 0.925, AUC = 0.978) when compared to alternative approaches, such as mixed features, supervised features, unsupervised/mixed/supervised feature fusion and selection, as well as patch-based convolutional neural network (CNN) feature extraction. We further validate the reproducibility and robustness of unsupervised feature extraction via stability and repeated splitting analysis, supporting its utility as a diagnostic aid in identifying CLL patients with histologic evidence of disease progression. The outcome of this study serves as proof of principle using an unsupervised machine learning scheme to enhance the diagnostic accuracy of the heterogeneous histology patterns that pathologists might not easily see.

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.

CellSpatialGraph: Integrate hierarchical phenotyping and graph modeling to characterize spatial architecture in tumor microenvironment on digital pathology.

We present CellSpatialGraph, an integrated clustering and graph-based framework, to investigate the cellular spatial structure. Due to the lack of a clear understanding of the cell subtypes in the tumor microenvironment, unsupervised learning is applied to uncover cell phenotypes. Then, we build local cell graphs, referred to as supercells, to model the cell-to-cell relationships at a local scale. After that, we apply clustering again to identify the subtypes of supercells. In the end, we build a global graph to summarize supercell-to-supercell interactions, from which we extract features to classify different disease subtypes.