Machine learning and radiomics for segmentation and classification of adnexal masses on ultrasound

Ultrasound-based models exist to support the classification of adnexal masses but are subjective and rely upon ultrasound expertise. We aimed to develop an end-to-end machine learning (ML) model capable of automating the classification of adnexal masses. In this retrospective study, transvaginal ultrasound scan images with linked diagnoses (ultrasound subjective assessment or histology) were extracted and segmented from Imperial College Healthcare, UK (ICH development dataset; n = 577 masses; 1444 images) and Morgagni-Pierantoni Hospital, Italy (MPH external dataset; n = 184 masses; 476 images). A segmentation and classification model was developed using convolutional neural networks and traditional radiomics features. Dice surface coefficient (DICE) was used to measure segmentation performance and area under the ROC curve (AUC), F1-score and recall for classification performance. The ICH and MPH datasets had a median age of 45 (IQR 35-60) and 48 (IQR 38-57) years old and consisted of 23.1% and 31.5% malignant cases, respectively. The best segmentation model achieved a DICE score of 0.85 ± 0.01, 0.88 ± 0.01 and 0.85 ± 0.01 in the ICH training, ICH validation and MPH test sets. The best classification model achieved a recall of 1.00 and F1-score of 0.88 (AUC:0.93), 0.94 (AUC:0.89) and 0.83 (AUC:0.90) in the ICH training, ICH validation and MPH test sets, respectively. We have developed an end-to-end radiomics-based model capable of adnexal mass segmentation and classification, with a comparable predictive performance (AUC 0.90) to the published performance of expert subjective assessment (gold standard), and current risk models. Further prospective evaluation of the classification performance of this ML model against existing methods is required.

Somatostatin Receptor Imaging with [18F]FET-βAG-TOCA PET/CT and [68Ga]Ga-DOTA-Peptide PET/CT in Patients with Neuroendocrine Tumors: A Prospective, Phase 2 Comparative Study

There is a clinical need for 18F-labeled somatostatin analogs for the imaging of neuroendocrine tumors (NET), given the limitations of using [68Ga]Ga-DOTA-peptides, particularly with regard to widespread accessibility. We have shown that [18F]fluoroethyl-triazole-[Tyr3]-octreotate ([18F]FET-βAG-TOCA) has favorable dosimetry and biodistribution. As a step toward clinical implementation, we conducted a prospective, noninferiority study of [18F]FET-βAG-TOCA PET/CT compared with [68Ga]Ga-DOTA- peptide PET/CT in patients with NET. Methods: Forty-five patients with histologically confirmed NET, grades 1 and 2, underwent PET/CT imaging with both [18F]FET-βAG-TOCA and [68Ga]Ga-peptide performed within a 6-mo window (median, 77 d; range, 6-180 d). Whole-body PET/CT was conducted 50 min after injection of 165 MBq of [18F]FET-βAG-TOCA. Tracer uptake was evaluated by comparing SUVmax and tumor-to-background ratios at both lesion and regional levels by 2 unblinded, experienced readers. A randomized, blinded reading of both scans was also then undertaken by 3 experienced readers, and consensus was assessed at a regional level. The ability of both tracers to visualize liver metastases was also assessed. Results: A total of 285 lesions were detected on both imaging modalities. An additional 13 tumor deposits were seen in 8 patients on [18F]FET-βAG-TOCA PET/CT, and [68Ga]Ga-DOTA-peptide PET/CT detected an additional 7 lesions in 5 patients. Excellent correlation in SUVmax was observed between both tracers (r = 0.91; P < 0.001). No difference was observed between median SUVmax across regions, except in the liver, where the median tumor-to-background ratio of [18F]FET-βAG-TOCA was significantly lower than that of [68Ga]Ga-DOTA-peptide (2.5 ± 1.9 vs. 3.5 ± 2.3; P < 0.001). Conclusion: [18F]FET-βAG-TOCA was not inferior to [68Ga]Ga-DOTA-peptide in visualizing NET and may be considered in routine clinical practice given the longer half-life and availability of the cyclotron-produced fluorine radioisotope.

Deep representation learning of tissue metabolome and computed tomography annotates NSCLC classification and prognosis

The rich chemical information from tissue metabolomics provides a powerful means to elaborate tissue physiology or tumor characteristics at cellular and tumor microenvironment levels. However, the process of obtaining such information requires invasive biopsies, is costly, and can delay clinical patient management. Conversely, computed tomography (CT) is a clinical standard of care but does not intuitively harbor histological or prognostic information. Furthermore, the ability to embed metabolome information into CT to subsequently use the learned representation for classification or prognosis has yet to be described. This study develops a deep learning-based framework -- tissue-metabolomic-radiomic-CT (TMR-CT) by combining 48 paired CT images and tumor/normal tissue metabolite intensities to generate ten image embeddings to infer metabolite-derived representation from CT alone. In clinical NSCLC settings, we ascertain whether TMR-CT results in an enhanced feature generation model solving histology classification/prognosis tasks in an unseen international CT dataset of 742 patients. TMR-CT non-invasively determines histological classes - adenocarcinoma/squamous cell carcinoma with an F1-score = 0.78 and further asserts patients' prognosis with a c-index = 0.72, surpassing the performance of radiomics models and deep learning on single modality CT feature extraction. Additionally, our work shows the potential to generate informative biology-inspired CT-led features to explore connections between hard-to-obtain tissue metabolic profiles and routine lesion-derived image data.

Bench to Bedside Development of [18F]Fluoromethyl-(1,2-2H4)choline ([18F]D4-FCH)

Malignant transformation is characterised by aberrant phospholipid metabolism of cancers, associated with the upregulation of choline kinase alpha (CHKα). Due to the metabolic instability of choline radiotracers and the increasing use of late-imaging protocols, we developed a more stable choline radiotracer, [18F]fluoromethyl-[1,2-2H4]choline ([18F]D4-FCH). [18F]D4-FCH has improved protection against choline oxidase, the key choline catabolic enzyme, via a 1H/2D isotope effect, together with fluorine substitution. Due to the promising mechanistic and safety profiles of [18F]D4-FCH in vitro and preclinically, the radiotracer has transitioned to clinical development. [18F]D4-FCH is a safe positron emission tomography (PET) tracer, with a favourable radiation dosimetry profile for clinical imaging. [18F]D4-FCH PET/CT in lung and prostate cancers has shown highly heterogeneous intratumoral distribution and large lesion variability. Treatment with abiraterone or enzalutamide in metastatic castrate-resistant prostate cancer patients elicited mixed responses on PET at 12-16 weeks despite predominantly stable radiological appearances. The sum of the weighted tumour-to-background ratios (TBRs-wsum) was associated with the duration of survival.

Development and Evaluation of Machine Learning in Whole-Body Magnetic Resonance Imaging for Detecting Metastases in Patients With Lung or Colon Cancer: A Diagnostic Test Accuracy Study

OBJECTIVES

Whole-body magnetic resonance imaging (WB-MRI) has been demonstrated to be efficient and cost-effective for cancer staging. The study aim was to develop a machine learning (ML) algorithm to improve radiologists' sensitivity and specificity for metastasis detection and reduce reading times.

MATERIALS AND METHODS

A retrospective analysis of 438 prospectively collected WB-MRI scans from multicenter Streamline studies (February 2013-September 2016) was undertaken. Disease sites were manually labeled using Streamline reference standard. Whole-body MRI scans were randomly allocated to training and testing sets. A model for malignant lesion detection was developed based on convolutional neural networks and a 2-stage training strategy. The final algorithm generated lesion probability heat maps. Using a concurrent reader paradigm, 25 radiologists (18 experienced, 7 inexperienced in WB-/MRI) were randomly allocated WB-MRI scans with or without ML support to detect malignant lesions over 2 or 3 reading rounds. Reads were undertaken in the setting of a diagnostic radiology reading room between November 2019 and March 2020. Reading times were recorded by a scribe. Prespecified analysis included sensitivity, specificity, interobserver agreement, and reading time of radiology readers to detect metastases with or without ML support. Reader performance for detection of the primary tumor was also evaluated.

RESULTS

Four hundred thirty-three evaluable WB-MRI scans were allocated to algorithm training (245) or radiology testing (50 patients with metastases, from primary 117 colon [n = 117] or lung [n = 71] cancer). Among a total 562 reads by experienced radiologists over 2 reading rounds, per-patient specificity was 86.2% (ML) and 87.7% (non-ML) (-1.5% difference; 95% confidence interval [CI], -6.4%, 3.5%; P = 0.39). Sensitivity was 66.0% (ML) and 70.0% (non-ML) (-4.0% difference; 95% CI, -13.5%, 5.5%; P = 0.344). Among 161 reads by inexperienced readers, per-patient specificity in both groups was 76.3% (0% difference; 95% CI, -15.0%, 15.0%; P = 0.613), with sensitivity of 73.3% (ML) and 60.0% (non-ML) (13.3% difference; 95% CI, -7.9%, 34.5%; P = 0.313). Per-site specificity was high (>90%) for all metastatic sites and experience levels. There was high sensitivity for the detection of primary tumors (lung cancer detection rate of 98.6% with and without ML [0.0% difference; 95% CI, -2.0%, 2.0%; P = 1.00], colon cancer detection rate of 89.0% with and 90.6% without ML [-1.7% difference; 95% CI, -5.6%, 2.2%; P = 0.65]). When combining all reads from rounds 1 and 2, reading times fell by 6.2% (95% CI, -22.8%, 10.0%) when using ML. Round 2 read-times fell by 32% (95% CI, 20.8%, 42.8%) compared with round 1. Within round 2, there was a significant decrease in read-time when using ML support, estimated as 286 seconds (or 11%) quicker ( P = 0.0281), using regression analysis to account for reader experience, read round, and tumor type. Interobserver variance suggests moderate agreement, Cohen κ = 0.64; 95% CI, 0.47, 0.81 (with ML), and Cohen κ = 0.66; 95% CI, 0.47, 0.81 (without ML).

CONCLUSIONS

There was no evidence of a significant difference in per-patient sensitivity and specificity for detecting metastases or the primary tumor using concurrent ML compared with standard WB-MRI. Radiology read-times with or without ML support fell for round 2 reads compared with round 1, suggesting that readers familiarized themselves with the study reading method. During the second reading round, there was a significant reduction in reading time when using ML support.

Radiomics-based decision support tool assists radiologists in small lung nodule classification and improves lung cancer early diagnosis

BACKGROUND

Methods to improve stratification of small (≤15 mm) lung nodules are needed. We aimed to develop a radiomics model to assist lung cancer diagnosis.

METHODS

Patients were retrospectively identified using health records from January 2007 to December 2018. The external test set was obtained from the national LIBRA study and a prospective Lung Cancer Screening programme. Radiomics features were extracted from multi-region CT segmentations using TexLab2.0. LASSO regression generated the 5-feature small nodule radiomics-predictive-vector (SN-RPV). K-means clustering was used to split patients into risk groups according to SN-RPV. Model performance was compared to 6 thoracic radiologists. SN-RPV and radiologist risk groups were combined to generate "Safety-Net" and "Early Diagnosis" decision-support tools.

RESULTS

In total, 810 patients with 990 nodules were included. The AUC for malignancy prediction was 0.85 (95% CI: 0.82-0.87), 0.78 (95% CI: 0.70-0.85) and 0.78 (95% CI: 0.59-0.92) for the training, test and external test datasets, respectively. The test set accuracy was 73% (95% CI: 65-81%) and resulted in 66.67% improvements in potentially missed [8/12] or delayed [6/9] cancers, compared to the radiologist with performance closest to the mean of six readers.

CONCLUSIONS

SN-RPV may provide net-benefit in terms of earlier cancer diagnosis.

Modulation of EGFR Activity by Molecularly Imprinted Polymer Nanoparticles Targeting Intracellular Epitopes

In recent years, molecularly imprinted polymer nanoparticles (nanoMIPs) have proven to be an attractive alternative to antibodies in diagnostic and therapeutic applications. However, several key questions remain: how suitable are intracellular epitopes as targets for nanoMIP binding? And to what extent can protein function be modulated via targeting specific epitopes? To investigate this, three extracellular and three intracellular epitopes of epidermal growth factor receptor (EGFR) were used as templates for the synthesis of nanoMIPs which were then used to treat cancer cells with different expression levels of EGFR. It was observed that nanoMIPs imprinted with epitopes from the intracellular kinase domain and the extracellular ligand binding domain of EGFR caused cells to form large foci of EGFR sequestered away from the cell surface, caused a reduction in autophosphorylation, and demonstrated effects on cell viability. Collectively, this suggests that intracellular domain-targeting nanoMIPs can be a potential new tool for cancer therapy.

Feasibility of [18F]fluoropivalate hybrid PET/MRI for imaging lower and higher grade glioma: a prospective first-in-patient pilot study

PURPOSE

MRI and PET are used in neuro-oncology for the detection and characterisation of lesions for malignancy to target surgical biopsy and to plan surgical resections or stereotactic radiosurgery. The critical role of short-chain fatty acids (SCFAs) in brain tumour biology has come to the forefront. The non-metabolised SCFA radiotracer, [18F]fluoropivalate (FPIA), shows low background signal in most tissues except eliminating organs and has appropriate human dosimetry. Tumour uptake of the radiotracer is, however, unknown. We investigated the uptake characteristics of FPIA in this pilot PET/MRI study.

METHODS

Ten adult glioma subjects were identified based on radiological features using standard-of-care MRI prior to any surgical intervention, with subsequent histopathological confirmation of glioma subtype and grade (lower-grade - LGG - and higher-grade - HGG - patients). FPIA was injected as an intravenous bolus injection (range 342-368 MBq), and dynamic PET and MRI data were acquired simultaneously over 66 min.

RESULTS

All patients tolerated the PET/MRI protocol. Three patients were reclassified following resection and histology. Tumour maximum standardised uptake value (SUVmax,60) increased in the order LGG (WHO grade 2) < HGG (WHO grade 3) < HGG (WHO grade 4). The net irreversible solute transfer, Ki, and influx rate constant, K1, were significantly higher in HGG (p < 0.05). Of the MRI variables studied, DCE-MRI-derived extravascular-and-extracellular volume fraction (ve) was high in tumours of WHO grade 4 compared with other grades (p < 0.05). SLC25A20 protein expression was higher in HGG compared with LGG.

CONCLUSION

Tumoural FPIA PET uptake is higher in HGG compared to LGG. This study supports further investigation of FPIA PET/MRI for brain tumour imaging in a larger patient population.

CLINICAL TRIAL REGISTRATION

Clinicaltrials.gov, NCT04097535.

Automated sulfur-[18F]fluoride exchange radiolabelling of a prostate specific membrane antigen (PSMA) targeted ligand using the GE FASTlab™ cassette-based platform

Sulfur-[18F]fluoride exchange radiochemistry is a rapid and convenient method for incorporating fluorine-18 into biologically active molecules. We report a fully automated radiolabelling procedure for the synthesis of a [18F]SO3F-bearing prostate specific membrane antigen (PSMA) targeted ligand ([18F]5) using the GE FASTLab™ cassette-based platform in a 25.0 ± 2.6% radiochemical yield (decay corrected). Uptake in vitro and in vivo correlated with PSMA expression, and the radioligand exhibited favourable biodistribution and pharmacokinetic profiles.

Clinical evaluation of atlas-based auto-segmentation in breast and nodal radiotherapy

OBJECTIVES

Accurate contouring of anatomical structures allows for high-precision radiotherapy planning, targeting the dose at treatment volumes and avoiding organs at risk. Manual contouring is time-consuming with significant user variability, whereas auto-segmentation (AS) has proven efficiency benefits but requires editing before treatment planning. This study investigated whether atlas-based AS (ABAS) accuracy improves with template atlas group size and character-specific atlas and test case selection.

METHODS AND MATERIALS

One clinician retrospectively contoured the breast, nodes, lung, heart, and brachial plexus on 100 CT scans, adhering to peer-reviewed guidelines. Atlases were clustered in group sizes, treatment positions, chest wall separations, and ASs created with Mirada software. The similarity of ASs compared to reference contours was described by the Jaccard similarity coefficient (JSC) and centroid distance variance (CDV).

RESULTS

Across group sizes, for all structures combined, the mean JSC was 0.6 (SD 0.3, p = .999). Across atlas-specific groups, 0.6 (SD 0.3, p = 1.000). The correlation between JSC and structure volume was weak in both scenarios (adjusted R2-0.007 and 0.185).Mean CDV was similar across groups but varied up to 1.2 cm for specific structures.

CONCLUSIONS

Character-specific atlas groups and test case selection did not improve accuracy outcomes. High-quality ASs were obtained from groups containing as few as ten atlases, subsequently simplifying the application of ABAS. CDV measures indicating auto-segmentation variations on the x, y, and z axes can be utilised to decide on the clinical relevance of variations and reduce AS editing.

ADVANCES IN KNOWLEDGE

High-quality ABASs can be obtained from as few as ten template atlases.Atlas and test case selection do not improve AS accuracy.Unlike well-known quantitative similarity indices, volume displacement metrics provide information on the location of segmentation variations, helping assessment of the clinical relevance of variations and reducing clinician editing. Volume displacement metrics combined with the qualitative measure of clinician assessment could reduce user variability.

Radiomics and artificial intelligence for precision medicine in lung cancer treatment

Lung cancer is the leading cause of cancer-related deaths worldwide. It exhibits, at the mesoscopic scale, phenotypic characteristics that are generally indiscernible to the human eye but can be captured non-invasively on medical imaging as radiomic features, which can form a high dimensional data space amenable to machine learning. Radiomic features can be harnessed and used in an artificial intelligence paradigm to risk stratify patients, and predict for histological and molecular findings, and clinical outcome measures, thereby facilitating precision medicine for improving patient care. Compared to tissue sampling-driven approaches, radiomics-based methods are superior for being non-invasive, reproducible, cheaper, and less susceptible to intra-tumoral heterogeneity. This review focuses on the application of radiomics, combined with artificial intelligence, for delivering precision medicine in lung cancer treatment, with discussion centered on pioneering and groundbreaking works, and future research directions in the area.

A Subset of Non-Small Cell Lung Cancer Patients Treated with Pemetrexed Show 18F-Fluorothymidine "Flare" on Positron Emission Tomography

Thymidylate synthase (TS) remains a major target for cancer therapy. TS inhibition elicits increases in DNA salvage pathway activity, detected as a transient compensatory "flare" in 3'-deoxy-3'-[18F]fluorothymidine positron emission tomography (18F-FLT PET). We determined the magnitude of the 18F-FLT flare in non-small cell lung cancer (NSCLC) patients treated with the antifolate pemetrexed in relation to clinical outcome.

METHOD

Twenty-one patients with advanced/metastatic non-small cell lung cancer (NSCLC) scheduled to receive palliative pemetrexed ± platinum-based chemotherapy underwent 18F-FLT PET at baseline and 4 h after initiating single-agent pemetrexed. Plasma deoxyuridine (dUrd) levels and thymidine kinase 1 (TK1) activity were measured before each scan. Patients were then treated with the combination therapy. The 18F-FLT PET variables were compared to RECIST 1.1 and overall survival (OS).

RESULTS

Nineteen patients had evaluable PET scans at both time points. A total of 32% (6/19) of patients showed 18F-FLT flares (>20% change in SUVmax-wsum). At the lesion level, only one patient had an FLT flare in all the lesions above (test-retest borders). The remaining had varied uptake. An 18F-FLT flare occurred in all lesions in 1 patient, while another patient had an 18F-FLT reduction in all lesions; 17 patients showed varied lesion uptake. All patients showed global TS inhibition reflected in plasma dUrd levels (p < 0.001) and 18F-FLT flares of TS-responsive normal tissues including small bowel and bone marrow (p = 0.004 each). Notably, 83% (5/6) of patients who exhibited 18F-FLT flares were also RECIST responders with a median OS of 31 m, unlike patients who did not exhibit 18F-FLT flares (15 m). Baseline plasma TK1 was prognostic of survival but its activity remained unchanged following treatment.

CONCLUSIONS

The better radiological response and longer survival observed in patients with an 18F-FLT flare suggest the efficacy of the tracer as an indicator of the early therapeutic response to pemetrexed in NSCLC.

Evaluation of [18F]AlF-EMP-105 for Molecular Imaging of C-Met

C-Met is a receptor tyrosine kinase that is overexpressed in a range of different cancer types, and has been identified as a potential biomarker for cancer imaging and therapy. Previously, a ⁶⁸Ga-labelled peptide, [⁶⁸Ga]Ga-EMP-100, has shown promise for imaging c-Met in renal cell carcinoma in humans. Herein, we report the synthesis and preliminary biological evaluation of an [¹⁸F]AlF-labelled analogue, [¹⁸F]AlF-EMP-105, for c-Met imaging by positron emission tomography. EMP-105 was radiolabelled using the aluminium-[¹⁸F]fluoride method with 46 ± 2% RCY and >95% RCP in 35-40 min. In vitro evaluation showed that [¹⁸F]AlF-EMP-105 has a high specificity for c-Met-expressing cells. Radioactive metabolite analysis at 5 and 30 min post-injection revealed that [¹⁸F]AlF-EMP-105 has good blood stability, but undergoes transformation-transchelation, defluorination or demetallation-in the liver and kidneys. PET imaging in non-tumour-bearing mice showed high radioactive accumulation in the kidneys, bladder and urine, demonstrating that the tracer is cleared predominantly as [¹⁸F]fluoride by the renal system. With its high specificity for c-Met expressing cells, [¹⁸F]AlF-EMP-105 shows promise as a potential diagnostic tool for imaging cancer.

Amplified therapeutic targets in high-grade serous ovarian carcinoma - a review of the literature with quantitative appraisal

High-grade serous ovarian carcinoma is a unique cancer characterised by universal TP53 mutations and widespread copy number alterations. These copy number alterations include deletion of tumour suppressors and amplification of driver oncogenes. Given their key oncogenic roles, amplified driver genes are often proposed as therapeutic targets. For example, development of anti-HER2 agents has been clinically successful in treatment of ERBB2-amplified tumours. A wide scope of preclinical work has since investigated numerous amplified genes as potential therapeutic targets in high-grade serous ovarian carcinoma. However, variable experimental procedures (e.g., choice of cell lines), ambiguous phenotypes or lack of validation hinders further clinical translation of many targets. In this review, we collate the genes proposed to be amplified therapeutic targets in high-grade serous ovarian carcinoma, and quantitatively appraise the evidence in support of each candidate gene. Forty-four genes are found to have evidence as amplified therapeutic targets; the five highest scoring genes are CCNE1, PAX8, URI1, PRKCI and FAL1. This review generates an up-to-date list of amplified therapeutic target candidates for further development and proposes comprehensive criteria to assist amplified therapeutic target discovery in the future.

Non-contrast CT synthesis using patch-based cycle-consistent generative adversarial network (Cycle-GAN) for radiomics and deep learning in the era of COVID-19

Handcrafted and deep learning (DL) radiomics are popular techniques used to develop computed tomography (CT) imaging-based artificial intelligence models for COVID-19 research. However, contrast heterogeneity from real-world datasets may impair model performance. Contrast-homogenous datasets present a potential solution. We developed a 3D patch-based cycle-consistent generative adversarial network (cycle-GAN) to synthesize non-contrast images from contrast CTs, as a data homogenization tool. We used a multi-centre dataset of 2078 scans from 1,650 patients with COVID-19. Few studies have previously evaluated GAN-generated images with handcrafted radiomics, DL and human assessment tasks. We evaluated the performance of our cycle-GAN with these three approaches. In a modified Turing-test, human experts identified synthetic vs acquired images, with a false positive rate of 67% and Fleiss' Kappa 0.06, attesting to the photorealism of the synthetic images. However, on testing performance of machine learning classifiers with radiomic features, performance decreased with use of synthetic images. Marked percentage difference was noted in feature values between pre- and post-GAN non-contrast images. With DL classification, deterioration in performance was observed with synthetic images. Our results show that whilst GANs can produce images sufficient to pass human assessment, caution is advised before GAN-synthesized images are used in medical imaging applications.

Tumor and local lymphoid tissue interaction determines prognosis in high-grade serous ovarian cancer

Tertiary lymphoid structure (TLS) is associated with prognosis in copy-number-driven tumors, including high-grade serous ovarian cancer (HGSOC), although the function of TLS and its interaction with copy-number alterations in HGSOC are not fully understood. In the current study, we confirm that TLS-high HGSOC patients show significantly better progression-free survival (PFS). We show that the presence of TLS in HGSOC tumors is associated with B cell maturation and cytotoxic tumor-specific T cell activation and proliferation. In addition, the copy-number loss of IL15 and CXCL10 may limit TLS formation in HGSOC; a list of genes that may dysregulate TLS function is also proposed. Last, a radiomics-based signature is developed to predict the presence of TLS, which independently predicts PFS in both HGSOC patients and immune checkpoint inhibitor (ICI)-treated non-small cell lung cancer (NSCLC) patients. Overall, we reveal that TLS coordinates intratumoral B cell and T cell response to HGSOC tumor, while the cancer genome evolves to counteract TLS formation and function.

An Integrated Clinical-MR Radiomics Model to Estimate Survival Time in Patients With Endometrial Cancer

BACKGROUND

Determination of survival time in women with endometrial cancer using clinical features remains imprecise. Features from MRI may improve the survival estimation allowing improved treatment planning.

PURPOSE

To identify clinical features and imaging signatures on T2-weighted MRI that can be used in an integrated model to estimate survival time for endometrial cancer subjects.

STUDY TYPE

Retrospective.

POPULATION

Four hundred thirteen patients with endometrial cancer as training (N = 330, 66.41 ± 11.42 years) and validation (N = 83, 67.60 ± 11.89 years) data and an independent set of 82 subjects as testing data (63.26 ± 12.38 years).

FIELD STRENGTH/SEQUENCE

1.5-T and 3-T scanners with sagittal T2-weighted spin echo sequence.

ASSESSMENT

Tumor regions were manually segmented on T2-weighted images. Features were extracted from segmented masks, and clinical variables including age, cancer histologic grade and risk score were included in a Cox proportional hazards (CPH) model. A group least absolute shrinkage and selection operator method was implemented to determine the model from the training and validation datasets.

STATISTICAL TESTS

A likelihood-ratio test and decision curve analysis were applied to compare the models. Concordance index (CI) and area under the receiver operating characteristic curves (AUCs) were calculated to assess the model.

RESULTS

Three radiomic features (two image intensity and volume features) and two clinical variables (age and cancer grade) were selected as predictors in the integrated model. The CI was 0.797 for the clinical model (includes clinical variables only) and 0.818 for the integrated model using training and validation datasets, the associated mean AUC value was 0.805 and 0.853. Using the testing dataset, the CI was 0.792 and 0.882, significantly different and the mean AUC was 0.624 and 0.727 for the clinical model and integrated model, respectively.

DATA CONCLUSION

The proposed CPH model with radiomic signatures may serve as a tool to improve estimated survival time in women with endometrial cancer.

EVIDENCE LEVEL

4 TECHNICAL EFFICACY: Stage 2.

A Novel Radiogenomics Biomarker for Predicting Treatment Response and Pneumotoxicity From Programmed Cell Death Protein or Ligand-1 Inhibition Immunotherapy in NSCLC

INTRODUCTION

Patient selection for checkpoint inhibitor immunotherapy is currently guided by programmed death-ligand 1 (PD-L1) expression obtained from immunohistochemical staining of tumor tissue samples. This approach is susceptible to limitations resulting from the dynamic and heterogeneous nature of cancer cells and the invasiveness of the tissue sampling procedure. To address these challenges, we developed a novel computed tomography (CT) radiomic-based signature for predicting disease response in patients with NSCLC undergoing programmed cell death protein 1 (PD-1) or PD-L1 checkpoint inhibitor immunotherapy.

METHODS

This retrospective study comprises a total of 194 patients with suitable CT scans out of 340. Using the radiomic features computed from segmented tumors on a discovery set of 85 contrast-enhanced chest CTs of patients diagnosed with having NSCLC and their CD274 count, RNA expression of the protein-encoding gene for PD-L1, as the response vector, we developed a composite radiomic signature, lung cancer immunotherapy-radiomics prediction vector (LCI-RPV). This was validated in two independent testing cohorts of 66 and 43 patients with NSCLC treated with PD-1 or PD-L1 inhibition immunotherapy, respectively.

RESULTS

LCI-RPV predicted PD-L1 positivity in both NSCLC testing cohorts (area under the curve [AUC] = 0.70, 95% confidence interval [CI]: 0.57-0.84 and AUC = 0.70, 95% CI: 0.46-0.94). In one cohort, it also demonstrated good prediction of cases with high PD-L1 expression exceeding key treatment thresholds (>50%: AUC = 0.72, 95% CI: 0.59-0.85 and >90%: AUC = 0.66, 95% CI: 0.45-0.88), the tumor's objective response to treatment at 3 months (AUC = 0.68, 95% CI: 0.52-0.85), and pneumonitis occurrence (AUC = 0.64, 95% CI: 0.48-0.80). LCI-RPV achieved statistically significant stratification of the patients into a high- and low-risk survival group (hazard ratio = 2.26, 95% CI: 1.21-4.24, p = 0.011 and hazard ratio = 2.45, 95% CI: 1.07-5.65, p = 0.035).

CONCLUSIONS

A CT radiomics-based signature developed from response vector CD274 can aid in evaluating patients' suitability for PD-1 or PD-L1 checkpoint inhibitor immunotherapy in NSCLC.

Prediction of Deep Myometrial Infiltration, Clinical Risk Category, Histological Type, and Lymphovascular Space Invasion in Women with Endometrial Cancer Based on Clinical and T2-Weighted MRI Radiomic Features

PURPOSE

To predict deep myometrial infiltration (DMI), clinical risk category, histological type, and lymphovascular space invasion (LVSI) in women with endometrial cancer using machine learning classification methods based on clinical and image signatures from T2-weighted MR images.

METHODS

A training dataset containing 413 patients and an independent testing dataset consisting of 82 cases were employed in this retrospective study. Manual segmentation of the whole tumor volume on sagittal T2-weighted MRI was performed. Clinical and radiomic features were extracted to predict: (i) DMI of endometrial cancer patients, (ii) endometrial cancer clinical high-risk level, (iii) histological subtype of tumor, and (iv) presence of LVSI. A classification model with different automatically selected hyperparameter values was created. The area under the curve (AUC) of the receiver operating characteristic (ROC) curve, F1 score, average recall, and average precision were calculated to evaluate different models.

RESULTS

Based on the independent external testing dataset, the AUCs for DMI, high-risk endometrial cancer, endometrial histological type, and LVSI classification were 0.79, 0.82, 0.91, and 0.85, respectively. The corresponding 95% confidence intervals (CI) of the AUCs were [0.69, 0.89], [0.75, 0.91], [0.83, 0.97], and [0.77, 0.93], respectively.

CONCLUSION

It is possible to classify endometrial cancer DMI, risk, histology type, and LVSI using different machine learning methods.

Somatostatin Receptor PET/MR Imaging of Inflammation in Patients With Large Vessel Vasculitis and Atherosclerosis

BACKGROUND

Assessing inflammatory disease activity in large vessel vasculitis (LVV) can be challenging by conventional measures.

OBJECTIVES

We aimed to investigate somatostatin receptor 2 (SST2) as a novel inflammation-specific molecular imaging target in LVV.

METHODS

In a prospective, observational cohort study, in vivo arterial SST2 expression was assessed by positron emission tomography/magnetic resonance imaging (PET/MRI) using 68Ga-DOTATATE and 18F-FET-βAG-TOCA. Ex vivo mapping of the imaging target was performed using immunofluorescence microscopy; imaging mass cytometry; and bulk, single-cell, and single-nucleus RNA sequencing.

RESULTS

Sixty-one participants (LVV: n = 27; recent atherosclerotic myocardial infarction of ≤2 weeks: n = 25; control subjects with an oncologic indication for imaging: n = 9) were included. Index vessel SST2 maximum tissue-to-blood ratio was 61.8% (P < 0.0001) higher in active/grumbling LVV than inactive LVV and 34.6% (P = 0.0002) higher than myocardial infarction, with good diagnostic accuracy (area under the curve: ≥0.86; P < 0.001 for both). Arterial SST2 signal was not elevated in any of the control subjects. SST2 PET/MRI was generally consistent with 18F-fluorodeoxyglucose PET/computed tomography imaging in LVV patients with contemporaneous clinical scans but with very low background signal in the brain and heart, allowing for unimpeded assessment of nearby coronary, myocardial, and intracranial artery involvement. Clinically effective treatment for LVV was associated with a 0.49 ± 0.24 (standard error of the mean [SEM]) (P = 0.04; 22.3%) reduction in the SST2 maximum tissue-to-blood ratio after 9.3 ± 3.2 months. SST2 expression was localized to macrophages, pericytes, and perivascular adipocytes in vasculitis specimens, with specific receptor binding confirmed by autoradiography. SSTR2-expressing macrophages coexpressed proinflammatory markers.

CONCLUSIONS

SST2 PET/MRI holds major promise for diagnosis and therapeutic monitoring in LVV. (PET Imaging of Giant Cell and Takayasu Arteritis [PITA], NCT04071691; Residual Inflammation and Plaque Progression Long-Term Evaluation [RIPPLE], NCT04073810).

Radiotheranostics in oncology: Making precision medicine possible

A quintessential setting for precision medicine, theranostics refers to a rapidly evolving field of medicine in which disease is diagnosed followed by treatment of disease-positive patients using tools for the therapy identical or similar to those used for the diagnosis. Against the backdrop of only-treat-when-visualized, the goal is a high therapeutic index with efficacy markedly surpassing toxicity. Oncology leads the way in theranostics innovation, where the approach has become possible with the identification of unique proteins and other factors selectively expressed in cancer versus healthy tissue, advances in imaging technology able to report these tissue factors, and major understanding of targeting chemicals and nanodevices together with methods to attach labels or warheads for imaging and therapy. Radiotheranostics-using radiopharmaceuticals-is becoming routine in patients with prostate cancer and neuroendocrine tumors who express the proteins PSMA (prostate-specific membrane antigen) and SSTR2 (somatostatin receptor 2), respectively, on their cancer. The palpable excitement in the field stems from the finding that a proportion of patients with large metastatic burden show complete and partial responses, and this outcome is catalyzing the search for more radiotheranostics approaches. Not every patient will benefit from radiotheranostics; but, for those who cross the target-detected line, the likelihood of response is very high.

A radiomics-based decision support tool improves lung cancer diagnosis in combination with the Herder score in large lung nodules

BACKGROUND

Large lung nodules (≥15 mm) have the highest risk of malignancy, and may exhibit important differences in phenotypic or clinical characteristics to their smaller counterparts. Existing risk models do not stratify large nodules well. We aimed to develop and validate an integrated segmentation and classification pipeline, incorporating deep-learning and traditional radiomics, to classify large lung nodules according to cancer risk.

METHODS

502 patients from five U.K. centres were recruited to the large-nodule arm of the retrospective LIBRA study between July 2020 and April 2022. 838 CT scans were used for model development, split into training and test sets (70% and 30% respectively). An nnUNet model was trained to automate lung nodule segmentation. A radiomics signature was developed to classify nodules according to malignancy risk. Performance of the radiomics model, termed the large-nodule radiomics predictive vector (LN-RPV), was compared to three radiologists and the Brock and Herder scores.

FINDINGS

499 patients had technically evaluable scans (mean age 69 ± 11, 257 men, 242 women). In the test set of 252 scans, the nnUNet achieved a DICE score of 0.86, and the LN-RPV achieved an AUC of 0.83 (95% CI 0.77-0.88) for malignancy classification. Performance was higher than the median radiologist (AUC 0.75 [95% CI 0.70-0.81], DeLong p = 0.03). LN-RPV was robust to auto-segmentation (ICC 0.94). For baseline solid nodules in the test set (117 patients), LN-RPV had an AUC of 0.87 (95% CI 0.80-0.93) compared to 0.67 (95% CI 0.55-0.76, DeLong p = 0.002) for the Brock score and 0.83 (95% CI 0.75-0.90, DeLong p = 0.4) for the Herder score. In the international external test set (n = 151), LN-RPV maintained an AUC of 0.75 (95% CI 0.63-0.85). 18 out of 22 (82%) malignant nodules in the Herder 10-70% category in the test set were identified as high risk by the decision-support tool, and may have been referred for earlier intervention.

INTERPRETATION

The model accurately segments and classifies large lung nodules, and may improve upon existing clinical models.

FUNDING

This project represents independent research funded by: 1) Royal Marsden Partners Cancer Alliance, 2) the Royal Marsden Cancer Charity, 3) the National Institute for Health Research (NIHR) Biomedical Research Centre at the Royal Marsden NHS Foundation Trust and The Institute of Cancer Research, London, 4) the National Institute for Health Research (NIHR) Biomedical Research Centre at Imperial College London, 5) Cancer Research UK (C309/A31316).

Assessing the In Vivo Biocompatibility of Molecularly Imprinted Polymer Nanoparticles

Molecularly imprinted polymer nanoparticles (nanoMIPs) are high affinity synthetic receptors which show promise as imaging and therapeutic agents. Comprehensive analysis of the in vivo behaviour of nanoMIPs must be performed before they can be considered for clinical applications. This work reports the solid-phase synthesis of nanoMIPs and an investigation of their biodistribution, clearance and cytotoxicity in a rat model following both intravenous and oral administration. These nanoMIPs were found in each harvested tissue type, including brain tissue, implying their ability to cross the blood-brain barrier. The nanoMIPs were cleared from the body via both faeces and urine. Furthermore, we describe an immunogenicity study in mice, demonstrating that nanoMIPs specific for a cell surface protein showed moderate adjuvant properties, whilst those imprinted for a scrambled peptide showed no such behaviour. Given their ability to access all tissue types and their relatively low cytotoxicity, these results pave the way for in vivo applications of nanoMIPs.

Design, synthesis, and evaluation of a novel PET imaging agent targeting lipofuscin in senescent cells

Promoting a senescent phenotype to suppress tumour progression may present an alternative strategy for treating cancer and encourages the development of positron emission tomography (PET) imaging biomarkers for assessing response to treatment. The accumulation of lipofuscin deposits in senescent cells is visualised using the pathology stain Sudan Black B (SBB) which is an emerging biomarker of senescence. We describe the design, synthesis and evaluation of [¹⁸F]fluoroethyltriazole-SBB ([¹⁸F]FET-SBB), a fluorine-18 radiolabelled derivative of SBB. The in vitro uptake of [¹⁸F]FET-SBB in a senescent cell line corelated with lipofuscin deposits; in vivo PET imaging and metabolite analysis confirm a favourable pharmacokinetic and metabolic profile for further studies of in vivo models of senescence.

Clinical translation of quantitative magnetic resonance imaging biomarkers - An overview and gap analysis of current practice

PURPOSE

This overview of the current landscape of quantitative magnetic resonance imaging biomarkers (qMR IBs) aims to support the standardisation of academic IBs to assist their translation to clinical practice.

METHODS

We used three complementary approaches to investigate qMR IB use and quality management practices within the UK: 1) a literature search of qMR and quality management terms during 2011-2015 and 2016-2020; 2) a database search for clinical research studies using qMR IBs during 2016-2020; and 3) a survey to ascertain the current availability and quality management practices for clinical MRI scanners and associated equipment at research institutions across the UK.

RESULTS

The analysis showed increased use of all qMR methods between the periods 2011-2015 and 2016-2020 and diffusion-tensor MRI and volumetry to be popular methods. However, the "translation ratio" of journal articles to clinical research studies was higher for qMR methods that have evidence of clinical translation via a commercial route, such as fat fraction and T2 mapping. The number of journal articles citing quality management terms doubled between the periods 2011-2015 and 2016-2020; although, its proportion relative to all journal articles only increased by 3.0%. The survey suggested that quality assurance (QA) and quality control (QC) of data acquisition procedures are under-reported in the literature and that QA/QC of acquired data/data analysis are under-developed and lack consistency between institutions.

CONCLUSIONS

We summarise current attempts to standardise and translate qMR IBs, and conclude by outlining the ideal quality management practices and providing a gap analysis between current practice and a metrological standard.

Development of a multi-task learning V-Net for pulmonary lobar segmentation on CT and application to diseased lungs

AIM

To develop a multi-task learning (MTL) V-Net for pulmonary lobar segmentation on computed tomography (CT) and application to diseased lungs.

MATERIALS AND METHODS

The described methodology utilises tracheobronchial tree information to enhance segmentation accuracy through the algorithm's spatial familiarity to define lobar extent more accurately. The method undertakes parallel segmentation of lobes and auxiliary tissues simultaneously by employing MTL in conjunction with V-Net-attention, a popular convolutional neural network in the imaging realm. Its performance was validated by an external dataset of patients with four distinct lung conditions: severe lung cancer, COVID-19 pneumonitis, collapsed lungs, and chronic obstructive pulmonary disease (COPD), even though the training data included none of these cases.

RESULTS

The following Dice scores were achieved on a per-segment basis: normal lungs 0.97, COPD 0.94, lung cancer 0.94, COVID-19 pneumonitis 0.94, and collapsed lung 0.92, all at p<0.05.

CONCLUSION

Despite severe abnormalities, the model provided good performance at segmenting lobes, demonstrating the benefit of tissue learning. The proposed model is poised for adoption in the clinical setting as a robust tool for radiologists and researchers to define the lobar distribution of lung diseases and aid in disease treatment planning.

Variational Bayes for high-dimensional proportional hazards models with applications within gene expression

MOTIVATION

Few Bayesian methods for analyzing high-dimensional sparse survival data provide scalable variable selection, effect estimation and uncertainty quantification. Such methods often either sacrifice uncertainty quantification by computing maximum a posteriori estimates, or quantify the uncertainty at high (unscalable) computational expense.

RESULTS

We bridge this gap and develop an interpretable and scalable Bayesian proportional hazards model for prediction and variable selection, referred to as sparse variational Bayes. Our method, based on a mean-field variational approximation, overcomes the high computational cost of Markov chain Monte Carlo, whilst retaining useful features, providing a posterior distribution for the parameters and offering a natural mechanism for variable selection via posterior inclusion probabilities. The performance of our proposed method is assessed via extensive simulations and compared against other state-of-the-art Bayesian variable selection methods, demonstrating comparable or better performance. Finally, we demonstrate how the proposed method can be used for variable selection on two transcriptomic datasets with censored survival outcomes, and how the uncertainty quantification offered by our method can be used to provide an interpretable assessment of patient risk.

AVAILABILITY AND IMPLEMENTATION

our method has been implemented as a freely available R package survival.svb (https://github.com/mkomod/survival.svb).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

[18F]Fluorothymidine(FLT)-PET imaging of thymidine kinase 1 pharmacodynamics in non-small cell lung cancer treated with pemetrexed

3070

Background: Imaging of tumor proliferation has been studied with FLT-PET in various tumor types including NSCLC. Pemetrexed inhibits thymidylate synthase(TS), dihydrofolate reductase (DHFR), and glycinamide ribonucleotide formyltransferase (GARFT). TS inhibition upregulates the thymidine salvage pathway including relocalisation of ENT1 to membrane and TK1 activation as a transient “flare” response. We hypothesise that this can be detected as an increase in FLT tumoral uptake that subsequently decreases with reduced proliferation. This study was conducted to assess FLT uptake as an early pharmacodynamics(PD) marker of pemetrexed response. Methods: This was an open-label imaging study in 21 patients with Stage 3/4 NSCLC treated with pemetrexed and platinum-based chemotherapy. Patients underwent FLT PET/CT scan at baseline and 4h after administration of pemetrexed. Platinum component of treatment was administered on the day after second FLT scan for cycle 1. Plasma for TK1 activity expression were collected before each scan time point and analysed by ELISA. Percentage change in standardized uptake value (%ΔSUV) was calculated as [SUV(PET2) – SUV(PET1)]/ SUV(PET1)*100. Treatment response calculated by RECIST 1. 1 and survival data were collected. Results: 17 patients had evaluable PET/CT scans for pemetrexed response. Median percentage difference for SUVmean and SUVmax in tumour lesions increased by 3% and 10.3% respectively. 5 patients showed homogeneous FLT flare at 4h after pemetrexed, 2 patients had decrease, 10 patients had heterogeneous FLT response (regardless of platinum doublet). There was no significant correlation between plasma TK1 activity and FLT flare. At 9 weeks, 4 patients had partial response, 9 stable disease and 4 progressive disease. Baseline and weighted average ΔSUVmax were not associated with survival. The 5 patients with FLT flare in all lesions showed a median OS of 31 months, unlike the group with heterogenous or decrease uptake(15 months). FLT uptake in bone marrow and small bowel significantly increased at 4h (t test p = 0.004, p = 0.004, respectively) indicating increased thymidine salvage activity. Early FLT uptake was not predictive for tumour RECIST response or OS. In multivariable cox regression analysis, pre-treatment TK1 activity, adjusted for performance status, smoking history and age, independently affected survival in this group (p = 0.011). Conclusions: Early FLT flare at 4h was seen in NSCLC post pemetrexed administration indicating activation of thymidine salvage pathway. Median overall survival of patients with an FLT flare response was more than twice longer than patients with mixed or no response. However, the small sample size lacked power to show statistical significance in the OS comparison. Further studies should evaluate this and the relationship to other prognostic variables in a larger cohort of patients. Clinical trial information: NCRI UK badge 9249.

The HERPET study: Imaging HER2 expression in breast cancer with the novel PET tracer [18F]GE-226, a first-in-patient study

3069

Background: Over-expression the human epidermal growth factor receptor-2 (HER2) is seen in 20% of breast cancers; this is an adverse prognostic factor and used to guide therapy selection. At present HER2 expression can only be determined using biopsy material using immunohistochemistry or fluorescence in situ hybridisation. Heterogeneous expression of HER2 is now being recognised as a cause of treatment resistance but is difficult to characterise. A non-invasive method for determining HER2 expression could have several advantages and help select appropriate therapy for patients. GE-226 is a novel radiolabelled GE-Affibody radioligand which binds to the HER2 receptor with high affinity at a different epitope than trastuzumab. Methods: Patients with locally advanced or metastatic breast cancer were recruited and scanned for 65 mins after iv injection of 200MBq of GE-226 (mean activity injected for each patient 202MBq (range 164-223MBq, mean radiochemical purity 94%) of radioligand, over one bed position for dynamic imaging, followed by a half-body scan. Blood sampling was used to measure metabolism of the tracer. Safety was assessed. HER2-extracellular domain (ECD) domain was measured in blood. Tumoural uptake was quantified by semi-quantitative and fquantitative parameters in HER2 positive and HER2 negative tumours. Patients had routine baseline FDG imaging. Results: Twenty patients completed the study. GE-226 scans were well tolerated. There were no serious adverse events. GE-226 was slowly metabolised into a single metabolite in the liver; 97% of parent remained at 60 minutes post injection (range 82-100). There was a significant difference in tumoural radioligand uptake between biopsy proven HER2 positive and HER2 negative tumoural patients as measured by SUVmean and SUVmax (p < 0.001). Comparing HER2 positive to HER2 negative cases, there was also a significant difference between tumour to normal tissue uptake ratios SUVmean. Heterogeneous uptake was observed in three patients, two with interlesional uptake variation and one with intralesional heterogeneity. Tumoural uptake increased over time. Normal physiological uptake in salivary glands and the thyroid gland was noted. GE-226 was able to differentiate between lymphadenopathy due to sarcoidosis and cancer in one patient and was superior to FDG which had shown widespread uptake in the benign and malignant nodes. Conclusions: [18F]GE-226 imaging is well tolerated and shows promise for imaging of HER2 positive breast cancer. Further studies with this agent are now planned. Clinical trial information: NCT03827317.

Iodo Silanes as Superior Substrates for the Solid Phase Synthesis of Molecularly Imprinted Polymer Nanoparticles

Current state-of-the-art techniques for the solid phase synthesis of molecularly imprinted polymer (MIP) nanoparticles typically rely on amino silanes for the immobilisation of template molecules prior to polymerisation. An investigation into commonly used amino silanes identified a number of problematic side reactions which negatively affect the purity and affinity of these polymers. Iodo silanes are presented as a superior alternative in a case study describing the synthesis of MIPs against epitopes of a common cancer biomarker, epidermal growth factor receptor (EGFR). The proposed iodo silane outperformed the amino silane by all metrics tested, showing high purity and specificity, and nanomolar affinity for the target peptide.

A comparison of machine learning methods for predicting recurrence and death after curative-intent radiotherapy for non-small cell lung cancer: Development and validation of multivariable clinical prediction models

BACKGROUND

Surveillance is universally recommended for non-small cell lung cancer (NSCLC) patients treated with curative-intent radiotherapy. High-quality evidence to inform optimal surveillance strategies is lacking. Machine learning demonstrates promise in accurate outcome prediction for a variety of health conditions. The purpose of this study was to utilise readily available patient, tumour, and treatment data to develop, validate and externally test machine learning models for predicting recurrence, recurrence-free survival (RFS) and overall survival (OS) at 2 years from treatment.

METHODS

A retrospective, multicentre study of patients receiving curative-intent radiotherapy for NSCLC was undertaken. A total of 657 patients from 5 hospitals were eligible for inclusion. Data pre-processing derived 34 features for predictive modelling. Combinations of 8 feature reduction methods and 10 machine learning classification algorithms were compared, producing risk-stratification models for predicting recurrence, RFS and OS. Models were compared with 10-fold cross validation and an external test set and benchmarked against TNM-stage and performance status. Youden Index was derived from validation set ROC curves to distinguish high and low risk groups and Kaplan-Meier analyses performed.

FINDINGS

Median follow-up time was 852 days. Parameters were well matched across training-validation and external test sets: Mean age was 73 and 71 respectively, and recurrence, RFS and OS rates at 2 years were 43% vs 34%, 54% vs 47% and 54% vs 47% respectively. The respective validation and test set AUCs were as follows: 1) RFS: 0·682 (0·575-0·788) and 0·681 (0·597-0·766), 2) Recurrence: 0·687 (0·582-0·793) and 0·722 (0·635-0·81), and 3) OS: 0·759 (0·663-0·855) and 0·717 (0·634-0·8). Our models were superior to TNM stage and performance status in predicting recurrence and OS.

INTERPRETATION

This robust and ready to use machine learning method, validated and externally tested, sets the stage for future clinical trials entailing quantitative personalised risk-stratification and surveillance following curative-intent radiotherapy for NSCLC.

FUNDING

A full list of funding bodies that contributed to this study can be found in the Acknowledgements section.

Integrating the OHIF Viewer into XNAT: Achievements, Challenges and Prospects for Quantitative Imaging Studies

Purpose: XNAT is an informatics software platform to support imaging research, particularly in the context of large, multicentre studies of the type that are essential to validate quantitative imaging biomarkers. XNAT provides import, archiving, processing and secure distribution facilities for image and related study data. Until recently, however, modern data visualisation and annotation tools were lacking on the XNAT platform. We describe the background to, and implementation of, an integration of the Open Health Imaging Foundation (OHIF) Viewer into the XNAT environment. We explain the challenges overcome and discuss future prospects for quantitative imaging studies. Materials and methods: The OHIF Viewer adopts an approach based on the DICOM web protocol. To allow operation in an XNAT environment, a data-routing methodology was developed to overcome the mismatch between the DICOM and XNAT information models and a custom viewer panel created to allow navigation within the viewer between different XNAT projects, subjects and imaging sessions. Modifications to the development environment were made to allow developers to test new code more easily against a live XNAT instance. Major new developments focused on the creation and storage of regions-of-interest (ROIs) and included: ROI creation and editing tools for both contour- and mask-based regions; a "smart CT" paintbrush tool; the integration of NVIDIA's Artificial Intelligence Assisted Annotation (AIAA); the ability to view surface meshes, fractional segmentation maps and image overlays; and a rapid image reader tool aimed at radiologists. We have incorporated the OHIF microscopy extension and, in parallel, introduced support for microscopy session types within XNAT for the first time. Results: Integration of the OHIF Viewer within XNAT has been highly successful and numerous additional and enhanced tools have been created in a programme started in 2017 that is still ongoing. The software has been downloaded more than 3700 times during the course of the development work reported here, demonstrating the impact of the work. Conclusions: The OHIF open-source, zero-footprint web viewer has been incorporated into the XNAT platform and is now used at many institutions worldwide. Further innovations are envisaged in the near future.

A practical guide to automating fluorine-18 PET radiochemistry using commercially available cassette-based platforms

This Tutorial Account aims to be a useful educational resource which describes how to automate fluorine-18 positron emission tomography (PET) radiochemistry using cassette-based automated radiosynthesis platforms.

Highly effective liquid and solid phase extraction methods to concentrate radioiodine isotopes for radioiodination chemistry

Radioactive iodine isotopes play a pivotal role in radiopharmaceuticals. Large-scale production of multi-patient dose of radioiodinated nuclear medicines requires high concentration of radioiodine. We demonstrate that tetrabutylammonium chloride and methyltrioctylamonium chloride are effective phase transfer reagents to concentrate iodide-124, iodide-125 and iodide-131 from the corresponding commercial water solutions. The resulting concentrated radioiodide, in the presence of either phase transfer reagent, does not hamper the chemical reactivity of aqueous radioiodide in the copper (II)-mediated one-pot three-component click chemistry to produce radioiodinated iodotriazoles.

Validation analysis of the novel imaging-based prognostic radiomic signature in patients undergoing primary surgery for advanced high-grade serous ovarian cancer (HGSOC)

BACKGROUND

Predictive models based on radiomics features are novel, highly promising approaches for gynaecological oncology. Here, we wish to assess the prognostic value of the newly discovered Radiomic Prognostic Vector (RPV) in an independent cohort of high-grade serous ovarian cancer (HGSOC) patients, treated within a Centre of Excellence, thus avoiding any bias in treatment quality.

METHODS

RPV was calculated using standardised algorithms following segmentation of routine preoperative imaging of patients (n = 323) who underwent upfront debulking surgery (01/2011-07/2018). RPV was correlated with operability, survival and adjusted for well-established prognostic factors (age, postoperative residual disease, stage), and compared to previous validation models.

RESULTS

The distribution of low, medium and high RPV scores was 54.2% (n = 175), 33.4% (n = 108) and 12.4% (n = 40) across the cohort, respectively. High RPV scores independently associated with significantly worse progression-free survival (PFS) (HR = 1.69; 95% CI:1.06-2.71; P = 0.038), even after adjusting for stage, age, performance status and residual disease. Moreover, lower RPV was significantly associated with total macroscopic tumour clearance (OR = 2.02; 95% CI:1.56-2.62; P = 0.00647).

CONCLUSIONS

RPV was validated to independently identify those HGSOC patients who will not be operated tumour-free in an optimal setting, and those who will relapse early despite complete tumour clearance upfront. Further prospective, multicentre trials with a translational aspect are warranted for the incorporation of this radiomics approach into clinical routine.

An overview of nuclear medicine research in the UK and the landscape for clinical adoption

BACKGROUND AND OBJECTIVES

Nuclear medicine contributes greatly to the clinical management of patients and experimental medicine. This report aims to (1) outline the current landscape of nuclear medicine research in the UK, including current facilities and recent or ongoing clinical studies and (2) provide information about the available pathways for clinical adoption and NHS funding (commissioning) of radiopharmaceuticals.

METHODS

Evidence was obtained through database searches for UK-based nuclear medicine clinical studies and by conducting a questionnaire-based survey of UK radiopharmaceutical production facilities. A recent history of clinical commissioning, either through recommendations from the National Institute for Health and Care Excellence (NICE) or through NHS specialised services commissioning, was compiled from publicly available documents and policies.

RESULTS

The collected data highlighted the UK's active nuclear medicine research community and recent investment in new facilities and upgrades. All commissioning routes favour radiopharmaceuticals that have marketing authorisation and since 2017 there has been a requirement to demonstrate both clinical and cost-effectiveness. Whilst radiopharmaceuticals for molecular radiotherapy are well suited to these commissioning pathways, diagnostic radiotracers have not historically been assessed in this manner.

CONCLUSIONS

We hope that by collating this information we will provide stimulus for future discussion and consensus statements around this topic.

Investigating CXCR4 expression of tumor cells and the vascular compartment: A multimodal approach

The C-X-C chemokine receptor 4 (CXCR4) is G protein-coupled receptor that upon binding to its cognate ligand, can lead to tumor progression. Several CXCR4-targeted therapies are currently under investigation, and with it comes the need for imaging agents capable of accurate depiction of CXCR4 for therapeutic stratification and monitoring. PET agents enjoy the most success, but more cost-effective and radiation-free approaches such as ultrasound (US) imaging could represent an attractive alternative. In this work, we developed a targeted microbubble (MB) for imaging of vascular CXCR4 expression in cancer. A CXCR4-targeted MB was developed through incorporation of the T140 peptide into the MB shell. Binding properties of the T140-MB and control, non-targeted MB (NT-MB) were evaluated in MDA-MB-231 cells where CXCR4 expression was knocked-down (via shRNA) through optical imaging, and in the lymphoma tumor models U2932 and SuDHL8 (high and low CXCR4 expression, respectively) by US imaging. PET imaging of [18F]MCFB, a tumor-penetrating CXCR4-targeted small molecule, was used to provide whole-tumor CXCR4 readouts. CXCR4 expression and microvessel density were performed by immunohistochemistry analysis and western blot. T140-MB were formed with similar properties to NT-MB and accumulated sensitively and specifically in cells according to their CXCR4 expression. In NOD SCID mice, T140-MB persisted longer in tumors than NT-MB, indicative of target interaction, but showed no difference between U2932 and SuDHL8. In contrast, PET imaging with [18F]MCFB showed a marked difference in tumor uptake at 40-60 min post-injection between the two tumor models (p<0.05). Ex vivo analysis revealed that the large differences in CXCR4 expression between the two models are not reflected in the vascular compartment, where the MB are restricted; in fact, microvessel density and CXCR4 expression in the vasculature was comparable between U2932 and SuDHL8 tumors. In conclusion, we successfully developed a T140-MB that can be used for imaging CXCR4 expression in the tumor vasculature.

A kit-based aluminium-[18F]fluoride approach to radiolabelled microbubbles

The production of 18F-labelled microbubbles (MBs) via the aluminium-[18F]fluoride ([18F]AlF) radiolabelling method and facile inverse-electron-demand Diels-Alder (IEDDA) 'click' chemistry is reported. An [18F]AlF-NODA-labelled tetrazine was synthesised in excellent radiochemical yield (>95% RCY) and efficiently conjugated to a trans-cyclooctene (TCO) functionalised phospholipid (40-50% RCY), which was incorporated into MBs (40-50% RCY). To demonstrate the potential of producing 18F-labelled MBs for clinical studies, we also describe a kit-based approach which is amenable for use in a hospital radiopharmacy setting.

Introduction to the National Cancer Imaging Translational Accelerator (NCITA): a UK-wide infrastructure for multicentre clinical translation of cancer imaging biomarkers

The National Cancer Imaging Translational Accelerator (NCITA) is creating a UK national coordinated infrastructure for accelerated translation of imaging biomarkers for clinical use. Through the development of standardised protocols, data integration tools and ongoing training programmes, NCITA provides a unique scalable infrastructure for imaging biomarker qualification using multicentre clinical studies.

Consideration of Metabolite Efflux in Radiolabelled Choline Kinetics

Hypoxia is a complex microenvironmental condition known to regulate choline kinase α (CHKA) activity and choline transport through transcription factor hypoxia-inducible factor-1α (HIF-1α) and, therefore, may confound the uptake of choline radiotracer [18F]fluoromethyl-[1,2-2H4]-choline ([18F]-D4-FCH). The aim of this study was to investigate how hypoxia affects the choline radiotracer dynamics. Three underlying mechanisms by which hypoxia could potentially alter the uptake of the choline radiotracer, [18F]-D4-FCH, were investigated: 18F-D4-FCH import, CHKA phosphorylation activity, and the efflux of [18F]-D4-FCH and its phosphorylated product [18F]-D4-FCHP. The effects of hypoxia on [18F]-D4-FCH uptake were studied in CHKA-overexpressing cell lines of prostate cancer, PC-3, and breast cancer MDA-MB-231 cells. The mechanisms of radiotracer efflux were assessed by the cell uptake and immunofluorescence in vitro and examined in vivo (n = 24). The mathematical modelling methodology was further developed to verify the efflux hypothesis using [18F]-D4-FCH dynamic PET scans from non-small cell lung cancer (NSCLC) patients (n = 17). We report a novel finding involving the export of phosphorylated [18F]-D4-FCH and [18F]-D4-FCHP via HIF-1α-responsive efflux transporters, including ABCB4, when the HIF-1α level is augmented. This is supported by a graphical analysis of human data with a compartmental model (M2T6k + k5) that accounts for the efflux. Hypoxia/HIF-1α increases the efflux of phosphorylated radiolabelled choline species, thus supporting the consideration of efflux in the modelling of radiotracer dynamics.

Caspase-Based PET for Evaluating Pro-Apoptotic Treatments in a Tuberculosis Mouse Model

PURPOSE

Despite recent advances in antimicrobial treatments, tuberculosis (TB) remains a major global health threat. Mycobacterium tuberculosis proliferates in macrophages, preventing apoptosis by inducing anti-apoptotic proteins leading to necrosis of the infected cells. Necrosis then leads to increased tissue destruction, reducing the penetration of antimicrobials and immune cells to the areas where they are needed most. Pro-apoptotic drugs could be used as host-directed therapies in TB to improve antimicrobial treatments and patient outcomes.

PROCEDURE

We evaluated [¹⁸F]-ICMT-11, a caspase-3/7-specific positron emission tomography (PET) radiotracer, in macrophage cell cultures and in an animal model of pulmonary TB that closely resembles human disease.

RESULTS

Cells infected with M. tuberculosis and treated with cisplatin accumulated [¹⁸F]-ICMT-11 at significantly higher levels compared with that of controls, which correlated with levels of caspase-3/7 activity. Infected mice treated with cisplatin with increased caspase-3/7 activity also had a higher [¹⁸F]-ICMT-11 PET signal compared with that of untreated infected animals.

CONCLUSIONS

[¹⁸F]-ICMT-11 PET could be used as a noninvasive approach to measure intralesional pro-apoptotic responses in situ in pulmonary TB models and support the development of pro-apoptotic host-directed therapies for TB.

Novel Non-Congeneric Derivatives of the Choline Kinase Alpha Inhibitor ICL-CCIC-0019

Choline kinase alpha (CHKA) is a promising target for the development of cancer therapeutics. We have previously reported ICL-CCIC-0019, a potent CHKA inhibitor with high cellular activity but with some unfavorable pharmacological properties. In this work, we present an active analogue of ICL-CCIC-0019 bearing a piperazine handle (CK146) to facilitate further structural elaboration of the pharmacophore and thus improve the biological profile. Two different strategies were evaluated in this study: (1) a prodrug approach whereby selective CHKA inhibition could be achieved through modulating the activity of CK146, via the incorporation of an ε-(Ac) Lys motif, cleavable by elevated levels of histone deacetylase (HDAC) and cathepsin L (CTSL) in tumour cells; (2) a prostate-specific membrane antigen (PSMA) receptor targeted delivery strategy. Prodrug (CK145) and PSMA-targeted (CK147) derivatives were successfully synthesized and evaluated in vitro. While the exploitation of CK146 in those two strategies did not deliver the expected results, important and informative structure-activity relationships were observed and have been reported.

Effective Detection and Monitoring of Glioma Using [18F]FPIA PET Imaging

BACKGROUND

Reprogrammed cellular metabolism is a cancer hallmark. In addition to increased glycolysis, the oxidation of acetate in the citric acid cycle is another common metabolic phenotype. We have recently developed a novel fluorine-18-labelled trimethylacetate-based radiotracer, [18F]fluoro-pivalic acid ([18F]FPIA), for imaging the transcellular flux of short-chain fatty acids, and investigated whether this radiotracer can be used for the detection of glioma growth.

METHODS

We evaluated the potential of [18F]FPIA PET to monitor tumor growth in orthotopic patient-derived (HSJD-GBM-001) and cell line-derived (U87, LN229) glioma xenografts, and also included [18F]FDG PET for comparison. We assessed proliferation (Ki-67) and the expression of lipid metabolism and transport proteins (CPT1, SLC22A2, SLC22A5, SLC25A20) by immunohistochemistry, along with etomoxir treatment to provide insights into [18F]FPIA uptake.

RESULTS

Longitudinal PET imaging showed gradual increase in [18F]FPIA uptake in orthotopic glioma models with disease progression (p < 0.0001), and high tumor-to-brain contrast compared to [18F]FDG (p < 0.0001). [18F]FPIA uptake correlated positively with Ki-67 (p < 0.01), SLC22A5 (p < 0.001) and SLC25A20 (p = 0.001), and negatively with CPT1 (p < 0.01) and SLC22A2 (p < 0.01). Etomoxir reduced [18F]FPIA uptake, which correlated with decreased Ki-67 (p < 0.05).

CONCLUSIONS

Our findings support the use of [18F]FPIA PET for the detection and longitudinal monitoring of glioma, showing a positive correlation with tumor proliferation, and suggest transcellular flux-mediated radiotracer uptake.

Detecting hypoxia in vitro using 18F-pretargeted IEDDA "click" chemistry in live cells

We have exemplified a pretargeted approach to interrogate hypoxia in live cells using radioactive bioorthogonal inverse electron demand Diels–Alder (IEDDA) “click” chemistry. Our novel ¹⁸F-tetrazine probe ([¹⁸F]FB-Tz) and 2-nitroimidazole-based TCO targeting molecule (8) showed statistically significant (P < 0.0001) uptake in hypoxic cells (ca. 90 %ID per mg) vs. normoxic cells (<10 %ID per mg) in a 60 min incubation of [¹⁸F]FB-Tz. This is the first time that an intracellularly targeted small-molecule for IEDDA “click” has been used in conjunction with a radioactive reporter molecule in live cells and may be a useful tool with far-reaching applicability for a variety of applications.

Bioorthogonal IEDDA “click” can interrogate intracellular hypoxia using a radioactive reporter molecule.

Radiolabelling an 18F biologic via facile IEDDA "click" chemistry on the GE FASTLab™ platform

The use of biologics in positron emission tomography (PET) imaging is an important area of radiopharmaceutical development and new automated methods are required to facilitate their production. We report an automated radiosynthesis method to produce a radiolabelled biologic via facile inverse electron demand Diels-Alder (IEDDA) "click" chemistry on a single GE FASTLab™ cassette. We exemplified the method by producing a fluorine-18 radiolabelled interleukin-2 (IL2) radioconjugate from a trans-cyclooctene (TCO) modified IL2 precursor. The radioconjugate was produced using a fully automated radiosynthesis on a single FASTLab™ cassette in a decay-corrected radiochemical yield (RCY, d.c.) of 19.8 ± 2.6% in 110 min (from start of synthesis); the molar activity was 132.3 ± 14.6 GBq μmol-1. The in vitro uptake of [18F]TTCO-IL2 correlated with the differential receptor expression (CD25, CD122, CD132) in PC3, NK-92 and activated human PBMCs. The automated method may be adapted for the radiosynthesis of any TCO-modified protein via IEDDA chemistry.

Synthesis and evaluation of 3'-[18F]fluorothymidine-5'-squaryl as a bioisostere of 3'-[18F]fluorothymidine-5'-monophosphate

The squaryl moiety has emerged as an important phosphate bioisostere with reportedly greater cell permeability. It has been used in the synthesis of several therapeutic drug molecules including nucleoside and nucleotide analogues but is yet to be evaluated in the context of positron emission tomography (PET) imaging. We have designed, synthesised and evaluated 3'-[18F]fluorothymidine-5'-squaryl ([18F]SqFLT) as a bioisostere to 3'-[18F]fluorothymidine-5'-monophosphate ([18F]FLTMP) for imaging thymidylate kinase (TMPK) activity. The overall radiochemical yield (RCY) was 6.7 ± 2.5% and radiochemical purity (RCP) was >90%. Biological evaluation in vitro showed low tracer uptake (<0.3% ID mg-1) but significantly discriminated between wildtype HCT116 and CRISPR/Cas9 generated TMPK knockdown HCT116shTMPK-. Evaluation of [18F]SqFLT in HCT116 and HCT116shTMPK- xenograft mouse models showed statistically significant differences in tumour uptake, but lacked an effective tissue retention mechanism, making the radiotracer in its current form unsuitable for PET imaging of proliferation.

Abstract PD6-09: Herpet study- PET imaging of HER2 expression in breast cancer using the novel Affibody tracer [18F]GE-226, a first in patient study

Background 20% of breast cancers have over-expression of the human epidermal growth factor receptor-2 (HER2), which is an adverse prognostic factor and used to guide therapy selection. At present HER2 expression can only be determined using biopsy material which is then analysed using immunohistochemistry or fluorescence in situ hybridisation. GE-226 is a radiolabelled Affibody® tracer which binds to the HER2 receptor with high affinity at a different epitope than trastuzumab. Heterogeneous expression does exist but the impact this has on treatment response has not been well assessed. A non-invasive method for determining HER2 expression could have several advantages and help select appropriate therapy for patients.

Trial DesignPatients with locally advanced or metastatic breast cancer were recruited and scanned for 65 mins after iv injection of 200MBq (mean activity injected for each patient 198 MBq (range 164-219MBq), mean radiochemical purity 94.6%) of tracer, with one dynamic bed position, and then a half-body scan was performed. Blood sampling was used to measure metabolism of the tracer. Safety was recorded. HER2-extracellular domain (ECD) domain was measured in blood. The original accrual target was 16 patients. Tumoural uptake was quantified by semi-quantitative and fully quantitative parameters in HER2 positive and HER2 negative tumours.

ResultsThirteen patients were recruited. Scans were well tolerated. There were no serious adverse events. GE-226 was metabolised into a single metabolite in the liver. 96.8 % parent remained at 60 minutes post injection. There was a significant difference between HER2 positive and HER2 negative tumoural uptake of tracer as measured by SUVmean and SUVmax (p&lt;0.05). Comparing HER2 positive to HER2 negative cases, there was also a significant difference between tumour to normal tissue uptake ratios (p&lt;0.05). Heterogeneous uptake was observed in the same patient. Tumoural uptake increased over time. Uptake in salivary glands and the thyroid gland was noted. In one patient GE-226 was able to differentiate between lymphadenopathy due to sarcoidosis and cancer and was superior to FDG which had shown widespread uptake in the same patient.

Conclusions[18F]GE-226 imaging is well tolerated and shows promise for imaging of HER2 positive breast cancer. Further studies with this agent are now planned.

Citation Format: Laura M. Kenny, Gosala S. Gopalakrishnan, Tara D. Barwick, Vijay Vaja, S. Hope McDevitt, Robert Punjani, Neva H. Patel, Rathi Ramakrishnan, Naina R. Patel, Stephen Johnston, Janine Mansi, Gary J. Cook, Fiona J. Gilbert, Franklin I. Aigbirhio, Duncan Hiscock, Eric O. Aboagye. Herpet study- PET imaging of HER2 expression in breast cancer using the novel Affibody tracer [18F]GE-226, a first in patient study [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD6-09.

Discovery of a biomarker candidate for surgical stratification in high-grade serous ovarian cancer

Background

Maximal effort cytoreductive surgery is associated with improved outcomes in advanced high-grade serous ovarian cancer (HGSOC). However, despite complete gross resection (CGR), there is a percentage of patients who will relapse and die early. The aim of this study is to identify potential candidate biomarkers to help personalise surgical radicality.

Methods

136 advanced HGSOC cases who underwent CGR were identified from three public transcriptomic datasets. Candidate prognostic biomarkers were discovered in this cohort by Cox regression analysis, and further validated by targeted RNA-sequencing in HGSOC cases from Imperial College Healthcare NHS Trust (n = 59), and a public dataset. Gene set enrichment analysis was performed to understand the biological significance of the candidate biomarker.

Results

We identified ALG5 as a prognostic biomarker for early tumour progression in advanced HGSOC despite CGR (HR = 2.42, 95% CI (1.57–3.75), p < 0.0001). The prognostic value of this new candidate biomarker was additionally confirmed in two independent datasets (HR = 1.60, 95% CI (1.03–2.49), p = 0.0368; HR = 3.08, 95% CI (1.07–8.81), p = 0.0365). Mechanistically, the oxidative phosphorylation was demonstrated as a potential biological pathway of ALG5-high expression in patients with early relapse (p < 0.001).

Conclusion

ALG5 has been identified as an independent prognostic biomarker for poor prognosis in advanced HGSOC patients despite CGR. This sets a promising platform for biomarker combinations and further validations towards future personalised surgical care.

The Oxford Classic Links Epithelial-to-Mesenchymal Transition to Immunosuppression in Poor Prognosis Ovarian Cancers

PURPOSE

Using RNA sequencing, we recently developed the 52-gene-based Oxford classifier of carcinoma of the ovary (Oxford Classic, OxC) for molecular stratification of serous ovarian cancers (SOCs) based on the molecular profiles of their cell of origin in the fallopian tube epithelium. Here, we developed a 52-gene NanoString panel for the OxC to test the robustness of the classifier.

EXPERIMENTAL DESIGN

We measured the expression of the 52 genes in an independent cohort of prospectively collected SOC samples (n = 150) from a homogenous cohort who were treated with maximal debulking surgery and chemotherapy. We performed data mining of published expression profiles of SOCs and validated the classifier results on tissue arrays comprising 137 SOCs.

RESULTS

We found evidence of profound nongenetic heterogeneity in SOCs. Approximately 20% of SOCs were classified as epithelial-to-mesenchymal transition-high (EMT-high) tumors, which were associated with poor survival. This was independent of established prognostic factors, such as tumor stage, tumor grade, and residual disease after surgery (HR, 3.3; P = 0.02). Mining expression data of 593 patients revealed a significant association between the EMT scores of tumors and the estimated fraction of alternatively activated macrophages (M2; P < 0.0001), suggesting a mechanistic link between immunosuppression and poor prognosis in EMT-high tumors.

CONCLUSIONS

The OxC-defined EMT-high SOCs carry particularly poor prognosis independent of established clinical parameters. These tumors are associated with high frequency of immunosuppressive macrophages, suggesting a potential therapeutic target to improve clinical outcome.

Transcriptional analysis of multiple ovarian cancer cohorts reveals prognostic and immunomodulatory consequences of ERV expression

BACKGROUND

Endogenous retroviruses (ERVs) play a role in a variety of biological processes, including embryogenesis and cancer. DNA methyltransferase inhibitors (DNMTi)-induced ERV expression triggers interferon responses in ovarian cancer cells via the viral sensing machinery. Baseline expression of ERVs also occurs in cancer cells, though this process is poorly understood and previously unexplored in epithelial ovarian cancer (EOC). Here, the prognostic and immunomodulatory consequences of baseline ERV expression was assessed in EOC.

METHODS

ERV expression was assessed using EOC transcriptional data from The Cancer Genome Atlas (TCGA) and from an independent cohort (Hammersmith Hospital, HH), as well as from untreated or DNMTi-treated EOC cell lines. Least absolute shrinkage and selection operator (LASSO) logistic regression defined an ERV expression score to predict patient prognosis. Immunohistochemistry (IHC) was conducted on the HH cohort. Combination of DNMTi treatment with γδ T cells was tested in vitro, using EOC cell lines and patient-derived tumor cells.

RESULTS

ERV expression was found to define clinically relevant subsets of EOC patients. An ERV prognostic score was successfully generated in TCGA and validated in the independent cohort. In EOC patients from this cohort, a high ERV score was associated with better survival (log-rank p=0.0009) and correlated with infiltration of CD8+PD1+T cells (r=0.46, p=0.0001). In the TCGA dataset, a higher ERV score was found in BRCA1/2 mutant tumors, compared to wild type (p=0.015), while a lower ERV score was found in CCNE1 amplified tumors, compared to wild type (p=0.019). In vitro, baseline ERV expression dictates the level of ERV induction in response to DNMTi. Manipulation of an ERV expression threshold by DNMTi resulted in improved EOC cell killing by cytotoxic immune cells.

CONCLUSIONS

These findings uncover the potential for baseline ERV expression to robustly inform EOC patient prognosis, influence tumor immune infiltration and affect antitumor immunity.