Resource requirements to accelerate clinical applications of next generation sequencing and radiomics: Workshop commentary and review.

The National Institutes of Health (NIH)/U.S. Food and Drug Administration (FDA) Joint Leadership Council Next-Generation Sequencing (NGS) and Radiomics Working Group (NGS&R WG) was formed by the NIH/FDA Joint Leadership Council to promote the development and validation of innovative NGS tests, radiomic tools, and associated data analysis and interpretation enhanced by artificial intelligence (AI) and machine-learning (ML) technologies. A two-day workshop was held on September 29-30, 2021 to convene members of the scientific community to discuss how to overcome the "ground truth" gap that has frequently been acknowledged as one of the limiting factors impeding high-quality research, development, validation, and regulatory science in these fields. This report provides a summary of the resource gaps identified by the WG and attendees, highlights existing resources and the ways they can potentially be leveraged to accelerate growth in these fields, and presents opportunities to support NGS and radiomic tool development and validation using technologies such as AI and ML.

Impact of prevalence and case distribution in lab-based diagnostic imaging studies.

We investigated effects of prevalence and case distribution on radiologist diagnostic performance as measured by area under the receiver operating characteristic curve (AUC) and sensitivity-specificity in lab-based reader studies evaluating imaging devices. Our retrospective reader studies compared full-field digital mammography (FFDM) to screen-film mammography (SFM) for women with dense breasts. Mammograms were acquired from the prospective Digital Mammographic Imaging Screening Trial. We performed five reader studies that differed in terms of cancer prevalence and the distribution of noncancers. Twenty radiologists participated in each reader study. Using split-plot study designs, we collected recall decisions and multilevel scores from the radiologists for calculating sensitivity, specificity, and AUC. Differences in reader-averaged AUCs slightly favored SFM over FFDM (biggest AUC difference: 0.047, SE = 0.023 , p = 0.047 ), where standard error accounts for reader and case variability. The differences were not significant at a level of 0.01 (0.05/5 reader studies). The differences in sensitivities and specificities were also indeterminate. Prevalence had little effect on AUC (largest difference: 0.02), whereas sensitivity increased and specificity decreased as prevalence increased. We found that AUC is robust to changes in prevalence, while radiologists were more aggressive with recall decisions as prevalence increased.

Covalently deposited dyes: a new chromogen paradigm that facilitates analysis of multiple biomarkers in situ.

Multiplexed analysis of multiple biomarkers in a tissue sample requires use of reporter dyes with specific spectral properties that enable discrimination of signals. Conventional chromogens with broad absorbance spectra, widely used in immunohistochemistry (IHC), offer limited utility for multiplexed detection. Many dyes with narrow absorbance spectra, eg rhodamines, fluoresceins, and cyanines, potentially useful for multiplexed detection are well-characterized; however, generation of a chromogenic reagent useful for IHC analysis has not been demonstrated. Studies reported herein demonstrate utility of tyramine-chemistry for synthesis of a wide variety of new chromogenic dye conjugates useful for multiplexed in situ analysis using conventional light microscopes. The dyes, useful individually or in blends to generate new colors, provide signal sensitivity and dynamic range similar to conventional DAB chromogen, while enabling analysis of co-localized biomarkers. It is anticipated that this new paradigm will enable generation of a wide variety of new chromogens, useful for both research and clinical biomarker analysis that will benefit clinicians and patients.

The significance of autoantibodies to DFS70/LEDGFp75 in health and disease: integrating basic science with clinical understanding.

Antinuclear autoantibodies (ANAs) displaying the nuclear dense fine speckled immunofluorescence (DFS-IIF) pattern in HEp-2 substrates are commonly observed in clinical laboratory referrals. They target the dense fine speckled autoantigen of 70 kD (DFS70), most commonly known as lens epithelium-derived growth factor p75 (LEDGFp75). Interesting features of these ANAs include their low frequency in patients with systemic autoimmune rheumatic diseases (SARD), elevated prevalence in apparently healthy individuals, IgG isotype, strong trend to occur as the only ANA specificity in serum, and occurrence in moderate to high titers. These autoantibodies have also been detected at varied frequencies in patients with diverse non-SARD inflammatory and malignant conditions such as atopic diseases, asthma, eye diseases, and prostate cancer. These observations have recently stimulated vigorous research on their clinical and biological significance. Some studies have suggested that they are natural, protective antibodies that could serve as biomarkers to exclude a SARD diagnosis. Other studies suggest that they might be pathogenic in certain contexts. The emerging role of DFS70/LEDGFp75 as a stress protein relevant to human acquired immunodeficiency syndrome, cancer, and inflammation also points to the possibility that these autoantibodies could be sensors of cellular stress and inflammation associated with environmental factors. In this comprehensive review, we integrate our current knowledge of the biology of DFS70/LEDGFp75 with the clinical understanding of its autoantibodies in the contexts of health and disease.

Computer-aided detection of endobronchial valves using volumetric CT.

RATIONALE AND OBJECTIVES: The ability to automatically detect and monitor implanted devices may serve an important role in patient care by aiding the evaluation of device and treatment efficacy. The purpose of this research was to develop a system for the automated detection of one-way endobronchial valves that were implanted for less invasive lung volume reduction. MATERIALS AND METHODS: Volumetric thin-section computed tomographic data was obtained for 194 subjects; 95 subjects implanted with 246 devices were used for system development and 99 subjects implanted with 354 devices were reserved for testing. The detection process consisted of preprocessing, pattern recognition based detection, and a final device selection. Following the preprocessing, a set of classifiers was trained using AdaBoost to discriminate true devices from false positives. The classifiers in the cascade used two simple features (either the mean or maximum attenuation) of a local region computed at multiple fixed landmarks relative to a template model of the valve. RESULTS: Free-response receiver-operating characteristic analysis was performed for the evaluation; the system could be set so the mean sensitivity was 96.5% with a mean of 0.18 false positives per subject. If knowledge of the number of implanted devices were incorporated, the sensitivity would be 96.9% with a mean of 0.061 false positives per subject; this corresponds to a total of 12 false negatives and six false positives for the 99 subjects in the test dataset. CONCLUSION: Software was developed for automated detection of endobronchial valves on volumetric computed tomography. The proposed device modeling and detection techniques may be applicable to other devices as well as useful for evaluation of treatment response.

Prevalence of tracheal collapse in an emphysema cohort as measured with end-expiration CT.

RATIONALE AND OBJECTIVES: To retrospectively investigate the prevalence of tracheal collapse in an emphysema cohort. The occurrence of a large degree of tracheal collapse may have important implications for the clinical management of respiratory symptoms and air trapping in patients with emphysema. MATERIALS AND METHODS: Paired full-inspiratory and end-expiratory thin-section volumetric computed tomographic scans were available for 1071 long-term smokers with clinically and physiologically confirmed emphysema. The percentage reduction in the cross-sectional tracheal luminal area from full-inspiration to end-expiration was automatically computed at 2.5-mm intervals along the centerline of the trachea using customized software. RESULTS: Maximal tracheal collapse did not follow a normal distribution in the emphysema cohort (P < .0001, skewness/kurtosis tests for normality); the median collapse was 18% (intraquartile range, 11%-30%). Statistically significant differences were found in the distribution of maximal collapse by gender (P < .005, Wilcoxon rank sum test). Overall, 10.5% of men and 17.1% of women showed evidence of tracheomalacia on the basis of the criterion of a reduction of 50% or greater in cross-sectional tracheal luminal area at end-expiration. CONCLUSION: This study offers insights into the prevalence of tracheal collapse in a cohort of patients with emphysema; future work is needed to determine the possible relationship between tracheal collapse and air trapping in subjects with emphysema.

Classification of parenchymal abnormality in scleroderma lung using a novel approach to denoise images collected via a multicenter study.

RATIONALE AND OBJECTIVES: Computerized classification techniques have been developed to offer accurate and robust pattern recognition in interstitial lung disease using texture features. However, these techniques still present challenges when analyzing computed tomographic (CT) image data from multiprotocols because of disparate acquisition protocols or from standardized, multicenter clinical trials because of noise variability. Our objective is to investigate the utility of denoising thin section CT image data to improve the classification of scleroderma disease patterns. The patterns are lung fibrosis (LF), groundglass (GG), honeycomb (HC), or normal lung (NL) within small regions of interest (ROIs). METHODS: High-resolution CT images were scanned in a multicenter clinical trial for the Scleroderma Lung Study. A thoracic radiologist contoured a training set (38 patients) consisting of 148 ROIs with 46 LF, 85 GG, 4 HC, and 13 NL patterns and contoured a test set (33 new patients) consisting of 132 ROIs with 44 LF, 72 GG, 4 HC, and 12 NL patterns. The corresponding CT slices of a contoured ROI were denoised using Aujol's mathematic partial differential equation algorithm. The algorithm's noise parameter was estimated as the standard deviation of grey-level signal (in Hounsfield units) in a homogeneous, non-lung region: the aorta. Within each contoured ROI, every pixel within a 4 x 4 neighborhood was sampled (4 x 4 grid sampling). All sampled pixels from a contoured ROI were assumed to be the same disease pattern as labeled by the radiologist. 5,690 pixels (3,009 LF, 1,994 GG, 348 HC, and 339 NL) and 5,045 pixels (2,665 LF, 1,753 GG, 291 HC, and 336 NL) were sampled in training and test sets, respectively. Next, 58 texture features from the original and denoised image were calculated for each pixel. Using a multinomial logistic model, subsets of features (one from original and another from denoised images) were selected to classify disease patterns. Finally, pixels were classified into disease patterns using a support vector machine procedure. RESULTS: From the training set, multinomial logistic model selected 45 features from the original images and 38 features from denoised images to classify disease patterns. Using the test set, the overall pixel classification rate by SVM increased from 87.8% to 89.5% with denoising. The specific classification rates (original/denoised) were 96.3/96.4% for LF, 88.8/89.4% for GG, 21.3/28.9% for HC, and 73.5/88.4% for NL. Denoising significantly improved the NL and overall classification rates (P = .037 and P = .047 respectively) at ROI level. CONCLUSIONS: Analyzing multicenter data using a denoising approach led to more parsimonious classification models with increasing accuracy. This approach offers a novel alternate classification strategy for heterogeneous technical and disease components. Furthermore, the model offers the potential to discriminate the multiple patterns of scleroderma disease correctly.

Automated classification of lung bronchovascular anatomy in CT using AdaBoost.

Lung CAD systems require the ability to classify a variety of pulmonary structures as part of the diagnostic process. The purpose of this work was to develop a methodology for fully automated voxel-by-voxel classification of airways, fissures, nodules, and vessels from chest CT images using a single feature set and classification method. Twenty-nine thin section CT scans were obtained from the Lung Image Database Consortium (LIDC). Multiple radiologists labeled voxels corresponding to the following structures: airways (trachea to 6th generation), major and minor lobar fissures, nodules, and vessels (hilum to peripheral), and normal lung parenchyma. The labeled data was used in conjunction with a supervised machine learning approach (AdaBoost) to train a set of ensemble classifiers. Each ensemble classifier was trained to detect voxels part of a specific structure (either airway, fissure, nodule, vessel, or parenchyma). The feature set consisted of voxel attenuation and a small number of features based on the eigenvalues of the Hessian matrix (used to differentiate structures by shape). When each ensemble classifier was composed of 20 weak classifiers, the AUC values for the airway, fissure, nodule, vessel, and parenchyma classifiers were 0.984+/-0.011, 0.949+/-0.009, 0.945+/-0.018, 0.953+/-0.016, and 0.931+/-0.015, respectively. The strong results suggest that this could be an effective input to higher-level anatomical based segmentation models with the potential to improve CAD performance.

Genome-wide transcriptional analysis of carboplatin response in chemosensitive and chemoresistant ovarian cancer cells.

We have recently described an ex vivo chemoresponse assay for determining chemosensitivity in primary cultures of human tumors. In this study, we have extended these experiments in an effort to correlate chemoresponse data with gene expression patterns at the level of transcription. Primary cultures of cells derived from ovarian carcinomas of individual patients (n=6) were characterized using the ChemoFx assay and classified as either carboplatin sensitive (n=3) or resistant (n=3). Three representative cultures of cells from each individual tumor were then subjected to Affymetrix gene chip analysis (n=18) using U95A human gene chip arrays. Data were analyzed using the dCHIP software package. We identified a significant number of genes whose expression patterns were altered between carboplatin chemosensitive and chemoresistant cells, in normal culture conditions and in the presence of carboplatin for either 2 or 72 hours. Among these differentially expressed genes, we found a significant proportion to be associated with apoptosis, cell-cell communication, cell adhesion, DNA repair, and cell proliferation. In general, the molecular phenotype displayed by chemoresistant cells was reflective of an extended life span in culture in the presence of carboplatin and the genes that define this phenotype are potential biomarkers for the prognostic management of ovarian cancer patients.

The ChemoFx assay: an ex vivo cell culture assay for predicting anticancer drug responses.

We provide a detailed description of the ChemoFx Assay, a phenotype-based cell culture assay for predicting anticancer drug responses in individual cancer patients. The ChemoFx Assay is based on the outgrowth and short-term primary culture of epithelial cells derived from pieces of solid tumor adenocarcinomas that are obtained at the time of tumor resection. Malignant epithelial cells are grown attached in wells of microtiter plates and treated with six escalating doses of chemotherapeutic drug. Using an operator-controlled automated image analysis system, cell kill is measured microscopically by counting the number of live cells remaining after dead cells have detached and are subsequently rinsed away. A dose-response graph is automatically generated by comparing the number of cells in drug-treated wells with those in control wells.

Expression of nucleolar-related proteins in porcine preimplantation embryos produced in vivo and in vitro.

The expression of nucleolar-related proteins was studied as an indirect marker of the ribosomal RNA (rRNA) gene activation in porcine embryos up to the blastocyst stage produced in vivo and in vitro. A group of the in vivo-developed embryos were cultured with alpha-amanitin to block the de novo embryonic mRNA transcription. Localization of proteins involved in the rRNA transcription (upstream binding factor [UBF], topoisomerase I, RNA polymerase I [RNA Pol I], and the RNA Pol I-associated factor PAF53) and processing (fibrillarin, nucleophosmin, and nucleolin) was assessed by immunocytochemistry and confocal laser-scanning microscopy, and mRNA expression was determined by semiquantitative reverse transcription-polymerase chain reaction (RT-PCR). These findings were correlated with ultrastructural data and autoradiography following 20-min [3H]uridine incubation. Additionally, expression of the pocket proteins pRb and p130, which are involved in cell-cycle regulation, was assessed by semiquantitative RT-PCR up to the blastocyst stage. Toward the end of third cell cycle, the nuclei in non-alpha-amanitin-treated, in vivo-produced embryos displayed different stages of transformation of the nuclear precursor bodies (NPBs) into fibrillogranular nucleoli associated with autoradiographic labeling. However, on culture with alpha-amanitin, NPBs were not transformed into a fibrillogranular nucleolus during this cell cycle, demonstrating that embryonic nucleogenesis requires de novo mRNA transcription. Moreover, immunolocalization of RNA Pol I, but not of UBF, and the mRNA expression of PAF53 and UBF were significantly reduced or absent after culture with alpha-amanitin, indicating that RNA Pol I, PAF53, and presumably, UBF are derived from de novo embryonic transcription. Embryonic genomic activation was delayed in porcine embryos produced in vitro compared to the in vivo-derived counterparts with respect to mRNAs encoding PAF53 and UBF. Moreover, differences existed in the mRNA expression patterns of pRb between in vivo- and in vitro-developed embryos. These findings show, to our knowledge for the first time, a nucleolus-related gene expression in the preimplantation porcine embryo, and they highlight the differences in quality between in vivo and in vitro-produced embryos.

Breast cancer--response rates to chemotherapeutic agents studied in vitro.

BACKGROUND: The biological efficacy of, and spectrum of action of, agents used in treatment of breast cancer are important issues in therapy planning. MATERIALS AND METHODS: Techniques used involve monolayer culture and a quantitative microtiter plate-based chemo-response assay. Precision Therapeutics' overall assessability rate is 90% for tumors of all types. In this study, 148 specimens derived from breast cancer were studied. Of these, 111 were additionally studied histopathologically. Ninety-two percent of the 111 specimens were confirmed to be epithelioid in nature and, thus, compatible with cells of breast cancer origin. RESULTS: In vitro chemo-response profiles indicated that individual agents stratified into groups, with cyclophosphamide and fluorouracil demonstrating responses of 69% and 57% respectively; doxorubicin, 45%; and docetaxel, paclitaxel and gemcitabine, 39, 27 and 36%, respectively. CONCLUSION: The spectrum of responsiveness of the individual agents was variable and not completely overlapping, as shown by the Venn diagrams.

Evidence for the isolation, growth, and characterization of malignant cells in primary cultures of human tumors.

Isolation and growth of malignant cells from solid tumors have often met with disappointing results. Consequently, we have developed a cell culture methodology based on ex vivo explantation of tumor tissue, with subsequent monolayer cell outgrowth. In an attempt to assess methods for detection of malignant cells in these cultures, we analyzed and compared the results of cytopathology, growth in soft agar, and detection of telomerase activity with those of standard immunohistochemistry (IHC) techniques for the detection of cytokeratins, tumor marker p53, and proliferation marker Ki-67. The sensitivity of detection of malignant cells was 85% (22/26) for cytopathological examination, 30% (3/10) for soft agar growth, and 100% (12/12) for detection of telomerase activity. From these data, we concluded that both cytopathological examination and assessment of telomerase activity contribute to the detection of malignant cells in primary cultures of human solid tumors, whereas growth in soft agar was not a good indicator of malignant cells. Although not specific for malignant cells per se, IHC detection for epithelial cell cytokeratins showed a high degree of sensitivity (100%, 23/23), whereas the sensitivity for detection of tumor marker p53 and proliferation marker Ki-67 was 30% (7/23) and 70% (16/23), respectively. These data also provide proof that malignant tumor cells, derived from a diverse number of human solid tumors, can be isolated and grown in primary cell culture.

Down-regulation of RNA helicase II/Gu results in the depletion of 18 and 28 S rRNAs in Xenopus oocyte.

Genetic manipulations have revealed the functions of RNA helicases in ribosomal RNA (rRNA) biogenesis in yeast. However, no report shows the role of an RNA helicase in rRNA formation in higher eukaryotes. This study reports the functional characterization of the frog homologue of nucleolar RNA helicase II/Gu (xGu or DDX21). Down-regulation of xGu in Xenopus laevis oocyte using an antisense oligodeoxynucleotide results in the depletion of 18 and 28 S rRNAs. The disappearance of 18 S rRNA is accompanied by an accumulation of 20 S, indicating that xGu is critical in the processing of 20 to 18 S rRNA. The degradation of 28 S rRNA into fragments smaller than 18 S is also associated with a specific decrease in the level of xGu protein. These effects are reversed in the presence of in vitro synthesized wild type xGu mRNA but not its helicase-deficient mutant form. Similar aberrant rRNA processing is observed when antibody against xGu is microinjected. The involvement of xGu in processing of rRNA is consistent with the localization of Gu protein to the granular and dense fibrillar components of PtK2 cell nucleoli by immunoelectron microscopy. Our results show that xGu is involved in the processing of 20 to 18 S rRNA and contributes to the stability of 28 S rRNA in Xenopus oocytes.

In vitro responses of ovarian cancers to platinums and taxanes.

We describe the in vitro patterns of response of explanted primary and recurrent ovarian cancers to platinum- and taxane-based chemotherapeutics. The chemoresponse assay utilizes cells that grow out from tumor fragments and are then challenged with varied concentrations of chemotherapeutic agents, coupled with a highly quantitative cell counting analysis system. The in vitro response rates for 268 primary cancer explants were 24% and 54% for carboplatin and cisplatin, respectively, and 31% and 25% for docetaxel and paclitaxel, respectively. Recurrent tumors presented lower rates of responsiveness, as expected. Furthermore, the chemotherapies worked on overlapping but distinct populations, even within the same class of drug, with 14% of the carboplatin-sensitive tumors being cisplatin-resistant and 59% of the cisplatin-sensitive tumors being carboplatin-resistant. These in vitro responses compare favorably to published in vivo clinical response rates. The current study serves to demonstrate how an in vitro predictive assay can be used as a surrogate for clinical therapeutic challenge.