Sequential deep learning image enhancement models improve diagnostic confidence, lesion detectability, and image reconstruction time in PET.

BACKGROUND: Investigate the potential benefits of sequential deployment of two deep learning (DL) algorithms namely DL-Enhancement (DLE) and DL-based time-of-flight (ToF) (DLT). DLE aims to enhance the rapidly reconstructed ordered-subset-expectation-maximisation algorithm (OSEM) images towards block-sequential-regularised-expectation-maximisation (BSREM) images, whereas DLT aims to improve the quality of BSREM images reconstructed without ToF. As the algorithms differ in their purpose, sequential application may allow benefits from each to be combined. 20 FDG PET-CT scans were performed on a Discovery 710 (D710) and 20 on Discovery MI (DMI; both GE HealthCare). PET data was reconstructed using five combinations of algorithms:1. ToF-BSREM, 2. ToF-OSEM + DLE, 3. OSEM + DLE + DLT, 4. ToF-OSEM + DLE + DLT, 5. ToF-BSREM + DLT. To assess image noise, 30 mm-diameter spherical VOIs were drawn in both lung and liver to measure standard deviation of voxels within the volume. In a blind clinical reading, two experienced readers rated the images on a five-point Likert scale based on lesion detectability, diagnostic confidence, and image quality. RESULTS: Applying DLE + DLT reduced noise whilst improving lesion detectability, diagnostic confidence, and image reconstruction time. ToF-OSEM + DLE + DLT reconstructions demonstrated an increase in lesion SUVmax of 28 ± 14% (average ± standard deviation) and 11 ± 5% for data acquired on the D710 and DMI, respectively. The same reconstruction scored highest in clinical readings for both lesion detectability and diagnostic confidence for D710. CONCLUSIONS: The combination of DLE and DLT increased diagnostic confidence and lesion detectability compared to ToF-BSREM images. As DLE + DLT used input OSEM images, and because DL inferencing was fast, there was a significant decrease in overall reconstruction time. This could have applications to total body PET.

The British nuclear medicine research strategy - the framework.

The British Nuclear Medicine Society (BNMS) has developed a Research Strategy framework led by the Research Champions of the BNMS and overseen by the BNMS Research and Innovation Committee. The objectives of the Research Strategy are to improve translation of cutting-edge nuclear medicine research from bench to bedside, the implementation of state-of-the-art multimodality technologies and to enhance multicentre radionuclide research in the UK. It strives to involve patients and the public in radionuclide research and to contribute to and work with the multi-professional national and international organisations involved in research with an ultimate aim to improve nuclear medicine services, and patients' outcomes and care.

On the use of solid 133Ba sources as surrogate for liquid 131I in SPECT/CT calibration: a European multi-centre evaluation.

INTRODUCTION: Commissioning, calibration, and quality control procedures for nuclear medicine imaging systems are typically performed using hollow containers filled with radionuclide solutions. This leads to multiple sources of uncertainty, many of which can be overcome by using traceable, sealed, long-lived surrogate sources containing a radionuclide of comparable energies and emission probabilities. This study presents the results of a quantitative SPECT/CT imaging comparison exercise performed within the MRTDosimetry consortium to assess the feasibility of using 133Ba as a surrogate for 131I imaging. MATERIALS AND METHODS: Two sets of four traceable 133Ba sources were produced at two National Metrology Institutes and encapsulated in 3D-printed cylinders (volume range 1.68-107.4 mL). Corresponding hollow cylinders to be filled with liquid 131I and a mounting baseplate for repeatable positioning within a Jaszczak phantom were also produced. A quantitative SPECT/CT imaging comparison exercise was conducted between seven members of the consortium (eight SPECT/CT systems from two major vendors) based on a standardised protocol. Each site had to perform three measurements with the two sets of 133Ba sources and liquid 131I. RESULTS: As anticipated, the 131I pseudo-image calibration factors (cps/MBq) were higher than those for 133Ba for all reconstructions and systems. A site-specific cross-calibration reduced the performance differences between both radionuclides with respect to a cross-calibration based on the ratio of emission probabilities from a median of 12-1.5%. The site-specific cross-calibration method also showed agreement between 133Ba and 131I for all cylinder volumes, which highlights the potential use of 133Ba sources to calculate recovery coefficients for partial volume correction. CONCLUSION: This comparison exercise demonstrated that traceable solid 133Ba sources can be used as surrogate for liquid 131I imaging. The use of solid surrogate sources could solve the radiation protection problem inherent in the preparation of phantoms with 131I liquid activity solutions as well as reduce the measurement uncertainties in the activity. This is particularly relevant for stability measurements, which have to be carried out at regular intervals.

Triple modality image reconstruction of PET data using SPECT, PET, CT information increases lesion uptake in images of patients treated with radioembolization with [Formula: see text] micro-spheres.

PURPOSE: Nuclear medicine imaging modalities like computed tomography (CT), single photon emission CT (SPECT) and positron emission tomography (PET) are employed in the field of theranostics to estimate and plan the dose delivered to tumors and the surrounding tissues and to monitor the effect of the therapy. However, therapeutic radionuclides often provide poor images, which translate to inaccurate treatment planning and inadequate monitoring images. Multimodality information can be exploited in the reconstruction to enhance image quality. Triple modality PET/SPECT/CT scanners are particularly useful in this context due to the easier registration process between images. In this study, we propose to include PET, SPECT and CT information in the reconstruction of PET data. The method is applied to Yttrium-90 ([Formula: see text]Y) data. METHODS: Data from a NEMA phantom filled with [Formula: see text]Y were used for validation. PET, SPECT and CT data from 10 patients treated with Selective Internal Radiation Therapy (SIRT) were used. Different combinations of prior images using the Hybrid kernelized expectation maximization were investigated in terms of VOI activity and noise suppression. RESULTS: Our results show that triple modality PET reconstruction provides significantly higher uptake when compared to the method used as standard in the hospital and OSEM. In particular, using CT-guided SPECT images, as guiding information in the PET reconstruction significantly increases uptake quantification on tumoral lesions. CONCLUSION: This work proposes the first triple modality reconstruction method and demonstrates up to 69% lesion uptake increase over standard methods with SIRT [Formula: see text]Y patient data. Promising results are expected for other radionuclide combination used in theranostic applications using PET and SPECT.

Variability of estimated glomerular filtration rate and 99m Tc-DTPA glomerular filtration rate: implications for a single time-point sampling regime.

BACKGROUND: This work aimed to determine the implications of the variability in estimated glomerular filtration rate (eGFR) for the prediction of measured GFR (mGFR) for selection of sampling time-point in single-sample 99m Tc-diethylene-triamine-pentaacetate (DTPA) mGFR. METHODS: Patient studies were used to compare eGFR and mGFR ( n = 282). The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration 2009 equation, from serum creatinine values measured in the laboratory ( n = 27) or using a point-of-care testing device ( n = 255). The mGFR was taken as the true value, and the root mean square error (RMS err ) in eGFR was calculated. Receiver operator characteristic curves were generated comparing the sensitivity and specificity of eGFR for the prediction of mGFR within the British Nuclear Medicine Society (BNMS) 2018 guideline ranges. RESULTS: The overall eGFR RMS err was 19.3 mL/min/1.73 m 2 . Use of eGFR to predict mGFR in the ranges specified in the BNMS 2018 guidelines (25-50; 50-70; 70-100; and >100) achieved the following specificity and sensitivity for each individual range (97%, 71%; 92%, 47%; 81%, 48%; and 74%, 90%). For the middle ranges (50-70 and 70-100) the sensitivity is very low, less than 50%; more studies are classified incorrectly on the basis of eGFR in these ranges than correctly. CONCLUSION: This work shows that serum creatinine eGFR is not sufficiently accurate to predict the optimum single-sample time-point for 99m Tc-DTPA mGFR prior to measurement. It is the recommendation of this study that a single sampling time-point should be chosen for studies eGFR > 40 ml/min/1.73 m 2 as opposed to the use of eGFR to determine the sampling time-point.

Reward insensitivity is associated with dopaminergic deficit in rapid eye movement sleep behaviour disorder.

Idiopathic rapid eye movement sleep behaviour disorder (iRBD) has now been established as an important marker of the prodromal stage of Parkinson's disease and related synucleinopathies. However, although dopamine transporter single photon emission computed tomography (SPECT) has been used to demonstrate the presence of nigro-striatal deficit in iRBD, quantifiable correlates of this are currently lacking. Sensitivity to rewarding stimuli is reduced in some people with Parkinson's disease, potentially contributing to aspects of the neuropsychiatric phenotype in these individuals. Furthermore, a role for dopaminergic degeneration is suggested by the fact that reward insensitivity can be improved by dopaminergic medications. Patients with iRBD present a unique opportunity to study the relationship between reward sensitivity and early dopaminergic deficit in the unmedicated state. Here, we investigate whether a non-invasive, objective measure of reward sensitivity might be a marker of dopaminergic status in prodromal Parkinson's disease by comparing with SPECT/CT measurement of dopaminergic loss in the basal ganglia. Striatal dopaminergic deficits in iRBD are associated with progression to Parkinsonian disorders. Therefore, identification of a clinically measurable correlate of this degenerative process might provide a basis for the development of novel risk stratification tools. Using a recently developed incentivized eye-tracking task, we quantified reward sensitivity in a cohort of 41 patients with iRBD and compared this with data from 40 patients with Parkinson's disease and 41 healthy controls. Patients with iRBD also underwent neuroimaging with dopamine transporter SPECT/CT. Overall, reward sensitivity, indexed by pupillary response to monetary incentives, was reduced in iRBD cases compared with controls and was not significantly different to that in patients with Parkinson's disease. However, in iRBD patients with normal dopamine transporter SPECT/CT imaging, reward sensitivity was not significantly different from healthy controls. Across all iRBD cases, a positive association was observed between reward sensitivity and dopaminergic SPECT/CT signal in the putamen. These findings demonstrate a direct relationship between dopaminergic deficit and reward sensitivity in patients with iRBD and suggest that measurement of pupillary responses could be of value in models of risk stratification and disease progression in these individuals.

Timing of hypoxia PET/CT imaging after 18F-fluoromisonidazole injection in non-small cell lung cancer patients.

Positron emission tomography (PET)/computed tomography (CT) using the radiotracer 18F-Fluoromisonidazole (FMISO) has been widely employed to image tumour hypoxia and is of interest to help develop novel hypoxia modifiers and guide radiation treatment planning. Yet, the optimal post-injection (p.i.) timing of hypoxic imaging remains questionable. Therefore, we investigated the correlation between hypoxia-related quantitative values in FMISO-PET acquired at 2 and 4 h p.i. in patients with non-small cell lung cancer (NSCLC). Patients with resectable NSCLC participated in the ATOM clinical trial (NCT02628080) which investigated the hypoxia modifying effects of atovaquone. Two-hour and four-hour FMISO PET/CT images acquired at baseline and pre-surgery visits (n = 58) were compared. Cohort 1 (n = 14) received atovaquone treatment, while cohort 2 (n = 15) did not. Spearman's rank correlation coefficients (ρ) assessed the relationship between hypoxia-related metrics, including standardised uptake value (SUV), tumour-to-blood ratio (TBR), and tumour hypoxic volume (HV) defined by voxels with TBR ≥ 1.4. As the primary imaging-related trial endpoint used to evaluate the action of atovaquone on tumour hypoxia in patients with NSCLC was change in tumour HV from baseline, this was also assessed in patients (n = 20) with sufficient baseline 2- and 4-h scan HV to reliably measure change (predefined as ≥ 1.5 mL). Tumours were divided into four subregions or distance categories: edge, outer, inner, and centre, using MATLAB. In tumours overall, strong correlation (P < 0.001) was observed for SUVmax ρ = 0.87, SUVmean ρ = 0.91, TBRmax ρ = 0.83 and TBRmean ρ = 0.81 between 2- and 4-h scans. Tumour HV was moderately correlated (P < 0.001) with ρ = 0.69 between 2- and 4-h scans. Yet, in tumour subregions, the correlation of HV decreased from the centre ρ = 0.71 to the edge ρ = 0.45 (P < 0.001). SUV, TBR, and HV values were consistently higher on 4-h scans than on 2-h scans, indicating better tracer-to-background contrast. For instance, for TBRmax, the mean, median, and interquartile range were 1.9, 1.7, and 1.6-2.0 2-h p.i., and 2.6, 2.4, and 2.0-3.0 4-h p.i., respectively. Our results support that FMISO-PET scans should be performed at 4 h p.i. to evaluate tumour hypoxia in NSCLC.Trial registration: ClinicalTrials.gov, NCT02628080. Registered 11/12/2015, https://clinicaltrials.gov/ct2/show/NCT02628080 .

The role of [18F]fluorodopa positron emission tomography in grading of gliomas.

PURPOSE: Gliomas are the most commonly occurring brain tumour in adults and there remains no cure for these tumours with treatment strategies being based on tumour grade. All treatment options aim to prolong survival, maintain quality of life and slow the inevitable progression from low-grade to high-grade. Despite imaging advancements, the only reliable method to grade a glioma is to perform a biopsy, and even this is fraught with errors associated with under grading. Positron emission tomography (PET) imaging with amino acid tracers such as [18F]fluorodopa (18F-FDOPA), [11C]methionine (11C-MET), [18F]fluoroethyltyrosine (18F-FET), and 18F-FDOPA are being increasingly used in the diagnosis and management of gliomas. METHODS: In this review we discuss the literature available on the ability of 18F-FDOPA-PET to distinguish low- from high-grade in newly diagnosed gliomas. RESULTS: In 2016 the Response Assessment in Neuro-Oncology (RANO) and European Association for Neuro-Oncology (EANO) published recommendations on the clinical use of PET imaging in gliomas. However, since these recommendations there have been a number of studies performed looking at whether 18F-FDOPA-PET can identify areas of high-grade transformation before the typical radiological features of transformation such as contrast enhancement are visible on standard magnetic resonance imaging (MRI). CONCLUSION: Larger studies are needed to validate 18F-FDOPA-PET as a non-invasive marker of glioma grade and prediction of tumour molecular characteristics which could guide decisions surrounding surgical resection.

Tailoring the sampling time of single-sample GFR measurement according to expected renal function: a multisite audit.

BACKGROUND: The 2018 BNMS Glomerular Filtration Rate (GFR) guidelines recommend a single-sample technique with the sampling time dictated by the expected renal function, but this is not known with any accuracy before the test. We aimed to assess whether the sampling regime suggested in the guidelines is optimal and determine the error in GFR result if the sample time is chosen incorrectly. We can then infer the degree of flexibility in the sampling regime. METHODS: Data from 6328 patients referred for GFR assessment at 6 different hospitals for a variety of indications were reviewed. The difference between the single-sample (Fleming) GFR result at each sample time and the slope-intercept GFR result at each hospital was calculated. A second dataset of 777 studies from one hospital with nine samples collected from 5 min to 8 h post-injection was analysed to provide a reference GFR to which the single-sample results were compared. RESULTS: Recommended single-sample times have been revised: for an expected GFR above 90 ml/min/1.73m2 a 2-h sample is recommended; between 50 and 90 ml/min/1.73m2 a 3-h sample is recommended; and between 30 and 50 ml/min/1.73m2 a 4-h sample is recommended. Root mean square error in single-sample GFR result compared with slope-intercept can be kept less than or equal to 3.30 ml/min/1.73m2 by following these recommendations. CONCLUSION: The results of this multisite study demonstrate a reassuringly wide range of sample times for an acceptably accurate single-sample GFR result. Modified recommended single-sample times have been proposed in line with the results, and a lookup table has been produced of rms errors across the full range of GFR results for the three sample times which can be used for error reporting of a mistimed sample.

Deep learning-based time-of-flight (ToF) image enhancement of non-ToF PET scans.

PURPOSE: To improve the quantitative accuracy and diagnostic confidence of PET images reconstructed without time-of-flight (ToF) using deep learning models trained for ToF image enhancement (DL-ToF). METHODS: A total of 273 [18F]-FDG PET scans were used, including data from 6 centres equipped with GE Discovery MI ToF scanners. PET data were reconstructed using the block-sequential-regularised-expectation-maximisation (BSREM) algorithm with and without ToF. The images were then split into training (n = 208), validation (n = 15), and testing (n = 50) sets. Three DL-ToF models were trained to transform non-ToF BSREM images to their target ToF images with different levels of DL-ToF strength (low, medium, high). The models were objectively evaluated using the testing set based on standardised uptake value (SUV) in 139 identified lesions, and in normal regions of liver and lungs. Three radiologists subjectively rated the models using testing sets based on lesion detectability, diagnostic confidence, and image noise/quality. RESULTS: The non-ToF, DL-ToF low, medium, and high methods resulted in - 28 ± 18, - 28 ± 19, - 8 ± 22, and 1.7 ± 24% differences (mean; SD) in the SUVmax for the lesions in testing set, compared to ToF-BSREM image. In background lung VOIs, the SUVmean differences were 7 ± 15, 0.6 ± 12, 1 ± 13, and 1 ± 11% respectively. In normal liver, SUVmean differences were 4 ± 5, 0.7 ± 4, 0.8 ± 4, and 0.1 ± 4%. Visual inspection showed that our DL-ToF improved feature sharpness and convergence towards ToF reconstruction. Blinded clinical readings of testing sets for diagnostic confidence (scale 0-5) showed that non-ToF, DL-ToF low, medium, and high, and ToF images scored 3.0, 3.0, 4.1, 3.8, and 3.5 respectively. For this set of images, DL-ToF medium therefore scored highest for diagnostic confidence. CONCLUSION: Deep learning-based image enhancement models may provide converged ToF-equivalent image quality without ToF reconstruction. In clinical scoring DL-ToF-enhanced non-ToF images (medium and high) on average scored as high as, or higher than, ToF images. The model is generalisable and hence, could be applied to non-ToF images from BGO-based PET/CT scanners.

Effects of Respiratory Motion on Y-90 PET Dosimetry for SIRT.

Respiratory motion degrades the quantification accuracy of PET imaging by blurring the radioactivity distribution. In the case of post-SIRT PET-CT verification imaging, respiratory motion can lead to inaccuracies in dosimetric measures. Using an anthropomorphic phantom filled with 90Y at a range of clinically relevant activities, together with a respiratory motion platform performing realistic motions (10-15 mm amplitude), we assessed the impact of respiratory motion on PET-derived post-SIRT dosimetry. Two PET scanners at two sites were included in the assessment. The phantom experiments showed that device-driven quiescent period respiratory motion correction improved the accuracy of the quantification with statistically significant increases in both the mean contrast recovery (+5%, p = 0.003) and the threshold activities corresponding to the dose to 80% of the volume of interest (+6%, p < 0.001). Although quiescent period gating also reduces the number of counts and hence increases the noise in the PET image, its use is encouraged where accurate quantification of the above metrics is desired.

Advances in PET/CT Technology: An Update.

This article reviews the current evolution and future directions in PET/CT technology focusing on three areas: time of flight, image reconstruction, and data-driven gating. Image reconstruction is considered with advances in point spread function modelling, Bayesian penalised likelihood reconstruction, and artificial intelligence approaches. Data-driven gating is examined with reference to respiratory motion, cardiac motion, and head motion. For each of these technological advancements, theory will be briefly discussed, benefits of their use in routine practice will be detailed and potential future developments will be discussed. Representative clinical cases will be presented, demonstrating the huge opportunities given to the PET community by hardware and software advances in PET technology when it comes to lesion detection, disease characterization, accurate quantitation and quicker scans. Through this review, hospitals are encouraged to embrace, evaluate and appropriately implement the wide range of new PET technologies that are available now or in the near future, for the improvement of patient care.

Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise.

PURPOSE: To enhance the image quality of oncology [18F]-FDG PET scans acquired in shorter times and reconstructed by faster algorithms using deep neural networks. METHODS: List-mode data from 277 [18F]-FDG PET/CT scans, from six centres using GE Discovery PET/CT scanners, were split into ¾-, ½- and »-duration scans. Full-duration datasets were reconstructed using the convergent block sequential regularised expectation maximisation (BSREM) algorithm. Short-duration datasets were reconstructed with the faster OSEM algorithm. The 277 examinations were divided into training (n = 237), validation (n = 15) and testing (n = 25) sets. Three deep learning enhancement (DLE) models were trained to map full and partial-duration OSEM images into their target full-duration BSREM images. In addition to standardised uptake value (SUV) evaluations in lesions, liver and lungs, two experienced radiologists scored the quality of testing set images and BSREM in a blinded clinical reading (175 series). RESULTS: OSEM reconstructions demonstrated up to 22% difference in lesion SUVmax, for different scan durations, compared to full-duration BSREM. Application of the DLE models reduced this difference significantly for full-, ¾- and ½-duration scans, while simultaneously reducing the noise in the liver. The clinical reading showed that the standard DLE model with full- or ¾-duration scans provided an image quality substantially comparable to full-duration scans with BSREM reconstruction, yet in a shorter reconstruction time. CONCLUSION: Deep learning-based image enhancement models may allow a reduction in scan time (or injected activity) by up to 50%, and can decrease reconstruction time to a third, while maintaining image quality.

Investigation of atovaquone-induced spatial changes in tumour hypoxia assessed by hypoxia PET/CT in non-small cell lung cancer patients.

BACKGROUND: Tumour hypoxia promotes an aggressive tumour phenotype and enhances resistance to anticancer treatments. Following the recent observation that the mitochondrial inhibitor atovaquone increases tumour oxygenation in NSCLC, we sought to assess whether atovaquone affects tumour subregions differently depending on their level of hypoxia. METHODS: Patients with resectable NSCLC participated in the ATOM trial (NCT02628080). Cohort 1 (n = 15) received atovaquone treatment, whilst cohort 2 (n = 15) did not. Hypoxia-related metrics, including change in mean tumour-to-blood ratio, tumour hypoxic volume, and fraction of hypoxic voxels, were assessed using hypoxia PET imaging. Tumours were divided into four subregions or distance categories: edge, outer, inner, and centre, using MATLAB. RESULTS: Atovaquone-induced reduction in tumour hypoxia mostly occurred in the inner and outer tumour subregions, and to a lesser extent in the centre subregion. Atovaquone did not seem to act in the edge subregion, which was the only tumour subregion that was non-hypoxic at baseline. Notably, the most intensely hypoxic tumour voxels, and therefore the most radiobiologically resistant areas, were subject to the most pronounced decrease in hypoxia in the different subregions. CONCLUSIONS: This study provides insights into the action of atovaquone in tumour subregions that help to better understand its role as a novel tumour radiosensitiser. TRIAL REGISTRATION: ClinicalTrials.gov, NCT0262808. Registered 11th December 2015, https://clinicaltrials.gov/ct2/show/NCT02628080.

A multicentre and multi-national evaluation of the accuracy of quantitative Lu-177 SPECT/CT imaging performed within the MRTDosimetry project.

PURPOSE: Patient-specific dosimetry is required to ensure the safety of molecular radiotherapy and to predict response. Dosimetry involves several steps, the first of which is the determination of the activity of the radiopharmaceutical taken up by an organ/lesion over time. As uncertainties propagate along each of the subsequent steps (integration of the time-activity curve, absorbed dose calculation), establishing a reliable activity quantification is essential. The MRTDosimetry project was a European initiative to bring together expertise in metrology and nuclear medicine research, with one main goal of standardizing quantitative 177Lu SPECT/CT imaging based on a calibration protocol developed and tested in a multicentre inter-comparison. This study presents the setup and results of this comparison exercise. METHODS: The inter-comparison included nine SPECT/CT systems. Each site performed a set of three measurements with the same setup (system, acquisition and reconstruction): (1) Determination of an image calibration for conversion from counts to activity concentration (large cylinder phantom), (2) determination of recovery coefficients for partial volume correction (IEC NEMA PET body phantom with sphere inserts), (3) validation of the established quantitative imaging setup using a 3D printed two-organ phantom (ICRP110-based kidney and spleen). In contrast to previous efforts, traceability of the activity measurement was required for each participant, and all participants were asked to calculate uncertainties for their SPECT-based activities. RESULTS: Similar combinations of imaging system and reconstruction lead to similar image calibration factors. The activity ratio results of the anthropomorphic phantom validation demonstrate significant harmonization of quantitative imaging performance between the sites with all sites falling within one standard deviation of the mean values for all inserts. Activity recovery was underestimated for total kidney, spleen, and kidney cortex, while it was overestimated for the medulla. CONCLUSION: This international comparison exercise demonstrates that harmonization of quantitative SPECT/CT is feasible when following very specific instructions of a dedicated calibration protocol, as developed within the MRTDosimetry project. While quantitative imaging performance demonstrates significant harmonization, an over- and underestimation of the activity recovery highlights the limitations of any partial volume correction in the presence of spill-in and spill-out between two adjacent volumes of interests.

Mitochondrial Inhibitor Atovaquone Increases Tumor Oxygenation and Inhibits Hypoxic Gene Expression in Patients with Non-Small Cell Lung Cancer.

PURPOSE: Tumor hypoxia fuels an aggressive tumor phenotype and confers resistance to anticancer treatments. We conducted a clinical trial to determine whether the antimalarial drug atovaquone, a known mitochondrial inhibitor, reduces hypoxia in non-small cell lung cancer (NSCLC). PATIENTS AND METHODS: Patients with NSCLC scheduled for surgery were recruited sequentially into two cohorts: cohort 1 received oral atovaquone at the standard clinical dose of 750 mg twice daily, while cohort 2 did not. Primary imaging endpoint was change in tumor hypoxic volume (HV) measured by hypoxia PET-CT. Intercohort comparison of hypoxia gene expression signatures using RNA sequencing from resected tumors was performed. RESULTS: Thirty patients were evaluable for hypoxia PET-CT analysis, 15 per cohort. Median treatment duration was 12 days. Eleven (73.3%) atovaquone-treated patients had meaningful HV reduction, with median change -28% [95% confidence interval (CI), -58.2 to -4.4]. In contrast, median change in untreated patients was +15.5% (95% CI, -6.5 to 35.5). Linear regression estimated the expected mean HV was 55% (95% CI, 24%-74%) lower in cohort 1 compared with cohort 2 (P = 0.004), adjusting for cohort, tumor volume, and baseline HV. A key pharmacodynamics endpoint was reduction in hypoxia-regulated genes, which were significantly downregulated in atovaquone-treated tumors. Data from multiple additional measures of tumor hypoxia and perfusion are presented. No atovaquone-related adverse events were reported. CONCLUSIONS: This is the first clinical evidence that targeting tumor mitochondrial metabolism can reduce hypoxia and produce relevant antitumor effects at the mRNA level. Repurposing atovaquone for this purpose may improve treatment outcomes for NSCLC.