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.

Fully automated deep-learning section-based muscle segmentation from CT images for sarcopenia assessment.

AIM: To develop a fully automated deep-learning-based approach to measure muscle area for assessing sarcopenia on standard-of-care computed tomography (CT) of the abdomen without any case exclusion criteria, for opportunistic screening for frailty. MATERIALS AND METHODS: This ethically approved retrospective study used publicly available and institutional unselected abdominal CT images (n=1,070 training, n=31 testing). The method consisted of two sequential steps: section detection from CT volume followed by muscle segmentation on single-section. Both stages used fully convolutional neural networks (FCNN), based on a UNet-like architecture. Input data consisted of CT volumes with a variety of fields of view, section thicknesses, occlusions, artefacts, and anatomical variations. Output consisted of segmented muscle area on a CT section at the L3 vertebral level. The muscle was segmented into erector spinae, psoas, and rectus abdominus muscle groups. Output was tested against expert manual segmentation. RESULTS: Threefold cross-validation was used to evaluate the model. Section detection cross-validation error was 1.41 ± 5.02 (in sections). Segmentation cross-validation Dice overlaps were 0.97 ± 0.02, 0.95 ± 0.04, and 0.94 ± 0.04 for erector spinae, psoas, and rectus abdominus, respectively, and 0.96 ± 0.02 for the combined muscle area, with R2 = 0.95/0.98 for muscle attenuation/area in 28/31 hold-out test cases. No statistical difference was found between the automated output and a second annotator. Fully automated processing took <1 second per CT examination. CONCLUSIONS: A FCNN pipeline accurately and efficiently automates muscle segmentation at the L3 vertebral level from unselected abdominal CT volumes, with no manual processing step. This approach is promising as a generalisable tool for opportunistic screening for frailty on standard-of-care CT.

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.

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.

Radio- and nano-chemistry of aqueous Ga(iii) ions anchored onto graphene oxide-modified complexes.

The gallium-68 radiolabelling of new functional graphene oxide composites is reported herein along with kinetic stability investigations of the radio-nanohybrids under different environments and insights into their surface characteristics by SEM and XPS. The present work highlights the potential of graphene oxides as nanocarriers for small molecules such as bis(thiosemicarbazonato) complexes to act as multifunctional platforms for rapid and effective radioimaging agent incorporation.

Machine learning in whole-body MRI: experiences and challenges from an applied study using multicentre data.

Machine learning is now being increasingly employed in radiology to assist with tasks such as automatic lesion detection, segmentation, and characterisation. We are currently involved in an National Institute of Health Research (NIHR)-funded project, which aims to develop machine learning methods to improve the diagnostic performance and reduce the radiology reading time of whole-body magnetic resonance imaging (MRI) scans, in patients being staged for cancer (MALIBO study). We describe here the main challenges we have encountered during the course of this project. Data quality and uniformity are the two most important data traits to be considered in clinical trials incorporating machine learning. Robust data pre-processing and machine learning pipelines have been employed in MALIBO, a task facilitated by the now freely available machine learning libraries and toolboxes. Another important consideration for achieving the desired clinical outcome in MALIBO, was to effectively host the resulting machine learning output, along with the clinical images, for reading in a clinical environment. Finally, a range of legal, ethical, and clinical acceptance issues should be considered when attempting to incorporate computer-assisting tools into clinical practice. The road from translating computational methods into potentially useful clinical tools involves an analytical, stepwise adaptation approach, as well as engagement of a multidisciplinary team.

Histogram analysis of apparent diffusion coefficient from whole-body diffusion-weighted MRI to predict early response to chemotherapy in patients with metastatic colorectal cancer: preliminary results.

AIM: To evaluate apparent diffusion coefficient (ADC) histogram analysis parameters, acquired from whole-body diffusion-weighted magnetic resonance imaging (DW-MRI), as very early predictors of response to chemotherapy in patients with metastatic colorectal cancer (mCRC). MATERIALS AND METHODS: This was a single-institution prospective study, approved by the West Midlands-South Birmingham research ethics committee. All patients gave fully informed consent prior to imaging. Sixteen patients with histologically confirmed mCRC were enrolled to the study and 11 were successfully scanned with whole-body DW-MRI before (baseline) and 10.8±2.7 days after commencing chemotherapy (follow-up). Therapy response was assessed by RECIST 1.1. Mean ADC and histogram parameters (skewness, kurtosis, 25th, 50th, and 75th percentiles) were compared between progressors and non-progressors at baseline and follow-up. Receiver operating characteristics (ROC) analysis was performed for the statistically significant parameters. Data from metastases were also compared to normative tissue data acquired from healthy volunteers. RESULTS: Three patients had progressive disease (progressors) and eight had partial response/stable disease (non-progressors). Mean, 25th, 50th, and 75th percentiles were significantly lower for progressors at baseline (p=0.012, 0.012, 0.012 and 0.025 respectively) with areas under the ROC curves (AUC)=0.58, 0.50, 0.58 and 0.63, respectively. Skewness and kurtosis were significantly lower for non-progressors at follow-up (p=0.001 and 0.003 respectively) with AUC=0.67 and 0.79 respectively. CONCLUSION: ADC histogram analysis shows potential in discriminating progressive from non-progressive disease in patients with mCRC, who underwent whole-body DW-MRI. The technique can potentially be tested as a response assessment methodology in larger trials.

Towards an MMP-2-activated molecular agent for cancer imaging.

Matrix metalloproteinases (MMPs) have been identified as biomarkers for cancer, offering prognostic potential; however, non-invasive detection protocols are currently lacking. Herein, we describe the synthesis of a DOTA-containing peptide sequence that can be radiolabelled easily with 68Gallium or can be incorporated with gadolinium for possible MRI applications with clear selectivity for MMP-2 over other members of the MMP family, giving MMP-2 selective cleavage of the labelled peptides.

Repeatability of quantitative 18F-FLT uptake measurements in solid tumors: an individual patient data multi-center meta-analysis.

INTRODUCTION: 3'-deoxy-3'-[18F]fluorothymidine (18F-FLT) positron emission tomography (PET) provides a non-invasive method to assess cellular proliferation and response to antitumor therapy. Quantitative 18F-FLT uptake metrics are being used for evaluation of proliferative response in investigational setting, however multi-center repeatability needs to be established. The aim of this study was to determine the repeatability of 18F-FLT tumor uptake metrics by re-analyzing individual patient data from previously published reports using the same tumor segmentation method and repeatability metrics across cohorts. METHODS: A systematic search in PubMed, EMBASE.com and the Cochrane Library from inception-October 2016 yielded five 18F-FLT repeatability cohorts in solid tumors. 18F-FLT avid lesions were delineated using a 50% isocontour adapted for local background on test and retest scans. SUVmax, SUVmean, SUVpeak, proliferative volume and total lesion uptake (TLU) were calculated. Repeatability was assessed using the repeatability coefficient (RC = 1.96 × SD of test-retest differences), linear regression analysis, and the intra-class correlation coefficient (ICC). The impact of different lesion selection criteria was also evaluated. RESULTS: Images from four cohorts containing 30 patients with 52 lesions were obtained and analyzed (ten in breast cancer, nine in head and neck squamous cell carcinoma, and 33 in non-small cell lung cancer patients). A good correlation was found between test-retest data for all 18F-FLT uptake metrics (R2 ≥ 0.93; ICC ≥ 0.96). Best repeatability was found for SUVpeak (RC: 23.1%), without significant differences in RC between different SUV metrics. Repeatability of proliferative volume (RC: 36.0%) and TLU (RC: 36.4%) was worse than SUV. Lesion selection methods based on SUVmax ≥ 4.0 improved the repeatability of volumetric metrics (RC: 26-28%), but did not affect the repeatability of SUV metrics. CONCLUSIONS: In multi-center studies, differences ≥ 25% in 18F-FLT SUV metrics likely represent a true change in tumor uptake. Larger differences are required for FLT metrics comprising volume estimates when no lesion selection criteria are applied.

Evaluation of 18F-fluorothymidine positron emission tomography ([18F]FLT-PET/CT) methodology in assessing early response to chemotherapy in patients with gastro-oesophageal cancer.

BACKGROUND: 3'-Deoxy-3'-[18F]fluorothymidine ([18F]FLT) PET has limited utility in abdominal imaging due to high physiological hepatic uptake of a tracer. We evaluated [18F]FLT-PET/CT combined with a temporal-intensity information-based voxel-clustering approach termed kinetic spatial filtering (KSF) to improve tumour visualisation in patients with locally advanced and metastatic gastro-oesophageal cancer and as a marker of early response to chemotherapy. Dynamic [18F]FLT-PET/CT data were collected before and 3 weeks post first cycle of chemotherapy. Changes in tumour [18F]FLT-PET/CT variables were determined. Response was determined on contrast-enhanced CT after three cycles of therapy using RECIST 1.1. RESULTS: Ten patients were included. Following application of the KSF, visual distinction of all oesophageal and/or gastric tumours was observed in [18F]FLT-PET images. Among the nine patients available for response evaluation (RECIST 1.1), three patients had responded (partial response) and six patients were non-responders (stable disease). There was a significant association between Ki-67 and all baseline [18F]FLT-PET parameters. Area under the curve (AUC) from 0 to 1 min was associated with treatment response. CONCLUSIONS: The results of this study indicate that application of the KSF allowed accurate visualisation of both primary and metastatic lesions following imaging with the proliferation marker, [18F]FLT-PET/CT. However, [18F]FLT-PET uptake parameters did not correlate with response. Instead, we observe significant changes in tracer delivery following chemotherapy suggesting that further [18F]FLT-PET/CT studies in this tumour type should be undertaken with caution.

AKT activation controls cell survival in response to HDAC6 inhibition.

HDAC6 is emerging as an important therapeutic target for cancer. We investigated mechanisms responsible for survival of tumor cells treated with a HDAC6 inhibitor. Expression of the 20 000 genes examined did not change following HDAC6 treatment in vivo. We found that HDAC6 inhibition led to an increase of AKT activation (P-AKT) in vitro, and genetic knockdown of HDAC6 phenocopied drug-induced AKT activation. The activation of AKT was not observed in PTEN null cells; otherwise, PTEN/PIK3CA expression per se did not predict HDAC6 inhibitor sensitivity. Interestingly, HDAC6 inhibitor treatment led to inactivating phosphorylation of PTEN (P-PTEN Ser380), which likely led to the increased P-AKT in cells that express PTEN. Synergy was observed with phosphatidylinositol 3'-kinases (PI3K) inhibitor treatment in vitro, accompanied by increased caspase 3/7 activity. Furthermore, combination of HDAC6 inhibitor with a PI3K inhibitor caused substantial tumor growth inhibition in vivo compared with either treatment alone, also detectable by Ki-67 immunostaining and (18)F-FLT positron emission tomography (PET). In aggregate AKT activation appears to be a key survival mechanism for HDAC6 inhibitor treatment. Our findings indicate that dual inhibition of HDAC6 and P-AKT may be necessary to substantially inhibit growth of solid tumors.

Significant metal enhanced fluorescence of Ag2S quantum dots in the second near-infrared window.

The amplification of light in NIR-II from Ag2S QDs via metal enhanced fluorescence (MEF) is reported for the first time. Significant fluorescence enhancement of over 100 times for Ag2S QDs deposited on Au-nanostructured arrays, paves the way for novel sensing and imaging applications based on Ag2S QDs, with improved detection sensitivity and contrast enhancement.

Design, synthesis and initial characterisation of a radiolabelled [(18)F]pyrimidoindolone probe for detecting activated caspase-3/7.

Evasion of apoptosis is one of the six initially proposed hallmarks of cancer, and as such, a method to detect apoptosis in a tumour would be of considerable interest in both clinical trials of new cancer therapeutics, as well as for routine patient management. Activation of caspase-3/7 is a key biomarker of cellular apoptosis. Herein we describe the design, synthesis and initial characterisation of the first pyrimidoindolone compound for detection of caspase-3/7 activation using positron emission tomography.

Mn-salen catalysed benzylic C-H activation for the synthesis of aryl [(18)F]CF3-containing PET probes.

The development of a Mn-salen complex catalysed oxidative benzylic fluorination of non-activated C-H bonds using [(18)F]fluoride is described for installation of [(18)F]CHRF, [(18)F]CR2F and particularly [(18)F]CF3 containing groups in the presence of other functional groups.

Epigenetic changes in gastroenteropancreatic neuroendocrine tumours.

An understanding of epigenetic drivers of tumorigenesis has developed rapidly during the last years. The identification of these changes including DNA methylation and histone modifications in gastroenteropancreatic neuroendocrine tumours (GEP-NETs) is a step forward in trying to define underlying biologic processes in this heterogeneous disease. The reversible nature of these changes represents a potential therapeutic target. We present an overview of the current knowledge of epigenetic alterations related to GEP-NETs, focusing on the influence and impact these changes have on pathogenesis and prognosis. The potential role of demethylating agents in the management of this patient population is discussed.

The anticancer gene ORCTL3 targets stearoyl-CoA desaturase-1 for tumour-specific apoptosis.

ORCTL3 is a member of a group of genes, the so-called anticancer genes, that cause tumour-specific cell death. We show that this activity is triggered in isogenic renal cells upon their transformation independently of the cells' proliferation status. For its cell death effect ORCTL3 targets the enzyme stearoyl-CoA desaturase-1 (SCD1) in fatty acid metabolism. This is caused by transmembrane domains 3 and 4, which are more efficacious in vitro than a low molecular weight drug against SCD1, and critically depend on their expression level. SCD1 is found upregulated upon renal cell transformation indicating that its activity, while not impacting proliferation, represents a critical bottleneck for tumourigenesis. An adenovirus expressing ORCTL3 leads to growth inhibition of renal tumours in vivo and to substantial destruction of patients' kidney tumour cells ex vivo. Our results indicate fatty acid metabolism as a target for tumour-specific apoptosis in renal tumours and suggest ORCTL3 as a means to accomplish this.

Parametric imaging of ¹8F-fluoro-3-deoxy-3-L-fluorothymidine PET data to investigate tumour heterogeneity.

PURPOSE: [(18)F]Fluoro-3'-deoxy-3'-L-fluorothymidine ([(18)F]FLT) is a tissue proliferation marker which has been widely validated as a tumour-specific imaging tracer for PET. [(18)F]FLT uptake in breast cancer is generally quantified at the region level or through first-order statistical descriptors (mean or maximum value), approaches that ignore the known complexity and heterogeneity of cancer tissues. Our aims were: (1) to validate a robust and reproducible voxel-wise approach to the quantification of [(18)F]FLT PET data in breast cancer patients, and (2) to exploit the entire distribution of the [(18)F]FLT retention estimates and their variability in the tumour region for the prediction of early treatment response. METHODS: The dataset was derived from 15 patients with stage II-IV breast cancer, scanned twice before chemotherapy and once 1 week after therapy. Using RECIST criteria (after 60 days) nine patients were categorized as responders or nonresponders to treatment. Kinetic modelling (compartmental modelling, Patlak analysis and spectral analysis with iterative filter), tissue-to-plasma ratio and standardized uptake value were applied at the voxel level. Test-retest estimates were used to assess reproducibility and reliability of the [(18)F]FLT uptake values before and after therapy for responder/nonresponder prediction. RESULTS: All the methods provided a measure of [(18)F]FLT uptake that was reliable and reproducible with ICC >0.94. Moreover, a very strong correlation was found among the methods (R (2) > 0.81). All the methods provided a limited number of outliers (<20 % in tumour), with the exception of compartmental modelling (>25 %) which was therefore excluded from the prediction analysis. Differences between before and after therapy in mean voxel-wise uptake in tumour did not allow a complete responder/nonresponder classification. In contrast, considering the full estimate distributions within the tumour (changes in median and mode between before and after therapy) improved therapy response for all the analysed methods. CONCLUSION: We showed that kinetic modelling (Patlak and spectral analysis with iterative filter) applied voxel-wise allows appropriate [(18)F]FLT uptake estimation in breast cancer with good reproducibility. Notably, this study indicated that a more comprehensive statistical investigation could improve tumour characterization and prediction of treatment response.

Development of a cyclin-dependent kinase inhibitor devoid of ABC transporter-dependent drug resistance.

BACKGROUND: Cyclin-dependent kinases (CDKs) control cell cycle progression, RNA transcription and apoptosis, making them attractive targets for anticancer drug development. Unfortunately, CDK inhibitors developed to date have demonstrated variable efficacy. METHODS: We generated drug-resistant cells by continuous low-dose exposure to a model pyrazolo[1,5-a]pyrimidine CDK inhibitor and investigated potential structural alterations for optimal efficacy. RESULTS: We identified induction of the ATP-binding cassette (ABC) transporters, ABCB1 and ABCG2, in resistant cells. Assessment of features involved in the ABC transporter substrate specificity from a compound library revealed high polar surface area (>100 Å(2)) as a key determinant of transporter interaction. We developed ICEC-0782 that preferentially inhibited CDK2, CDK7 and CDK9 in the nanomolar range. The compound inhibited phosphorylation of CDK substrates and downregulated the short-lived proteins, Mcl-1 and cyclin D1. ICEC-0782 induced G2/M arrest and apoptosis. The permeability and cytotoxicity of ICEC-0782 were unaffected by ABC transporter expression. Following daily oral dosing, the compound inhibited growth of human colon HCT-116 and human breast MCF7 tumour xenografts in vivo by 84% and 94%, respectively. CONCLUSION: We identified a promising pyrazolo[1,5-a]pyrimidine compound devoid of ABC transporter interaction, highly suitable for further preclinical and clinical evaluation for the treatment of cancer.

A novel small molecule hydroxamate preferentially inhibits HDAC6 activity and tumour growth.

BACKGROUND: This study investigates whether a histone deacetylase subtype 6 (HDAC6) inhibitor could be used in the treatment of solid tumours. METHODS: We evaluated the effect of a novel inhibitor, C1A, on HDAC6 biochemical activity and cell growth. We further examined potential of early noninvasive imaging of cell proliferation by [(18)F]fluorothymidine positron emission tomography ([(18)F]FLT-PET) to detect therapy response. RESULTS: C1A induced sustained acetylation of HDAC6 substrates, α-tubulin and HSP90, compared with current clinically approved HDAC inhibitor SAHA. C1A induced apoptosis and inhibited proliferation of a panel of human tumour cell lines from different origins in the low micromolar range. Systemic administration of the drug inhibited the growth of colon tumours in vivo by 78%. The drug showed restricted activity on gene expression with <0.065% of genes modulated during 24 h of treatment. C1A treatment reduced tumour [(18)F]FLT uptake by 1.7-fold at 48 h, suggesting that molecular imaging could provide value in future studies of this compound. CONCLUSION: C1A preferentially inhibits HDAC6 and modulates HDAC6 downstream targets leading to growth inhibition of a diverse set of cancer cell lines. This property together with the favourable pharmacokinetics and efficacy in vivo makes it a candidate for further pre-clinical and clinical development.

An open-label study of lapatinib in women with HER-2-negative early breast cancer: the lapatinib pre-surgical study (LPS study).

BACKGROUND: This phase II, open-label, multicentre study aimed to evaluate changes in cell proliferation and biomarkers, as well as efficacy of lapatinib in treatment-naïve patients with HER-2-negative primary breast cancer. PATIENTS AND METHODS: Patients received 1500 mg lapatinib for 28-42 days before surgery with repeat biopsies and measurements. The primary end point was inhibition of cell proliferation measured by Ki67; the secondary end points included clinical response, adverse events and changes in FOXO3a, FOXM1, p-AKT and HER-3. RESULTS: Overall, there was no significant reduction in Ki67 with treatment (assessment carried out in 28 of 31 subjects enrolled). However, four patients (14%) showed a reduction in Ki67 ≥50%. Four of 25 patients (16%) had a partial response to treatment judged by sequential ultrasound measurements. Response, in terms of either Ki67 or ultrasound, did not relate to changes in any biomarker assessed at baseline, including the estrogen receptor (ER) and epidermal growth factor receptor (EGFR). However, all four clinical responders were HER-3 positive, as were three of four Ki67 responders. CONCLUSIONS: Overall, a pre-surgical course of lapatinib monotherapy had little effect on this group of patients; however, in subsets of patients, especially those with HER-3-positive tumors, we observed either reduction in proliferation (Ki67) or tumor size; EGFR/ER status had no impact.