A platform-independent AI tumor lineage and site (ATLAS) classifier

Histopathologic diagnosis and classification of cancer plays a critical role in guiding treatment. Advances in next-generation sequencing have ushered in new complementary molecular frameworks. However, existing approaches do not independently assess both site-of-origin (e.g. prostate) and lineage (e.g. adenocarcinoma) and have minimal validation in metastatic disease, where classification is more difficult. Utilizing gradient-boosted machine learning, we developed ATLAS, a pair of separate AI Tumor Lineage and Site-of-origin models from RNA expression data on 8249 tumor samples. We assessed performance independently in 10,376 total tumor samples, including 1490 metastatic samples, achieving an accuracy of 91.4% for cancer site-of-origin and 97.1% for cancer lineage. High confidence predictions (encompassing the majority of cases) were accurate 98-99% of the time in both localized and remarkably even in metastatic samples. We also identified emergent properties of our lineage scores for tumor types on which the model was never trained (zero-shot learning). Adenocarcinoma/sarcoma lineage scores differentiated epithelioid from biphasic/sarcomatoid mesothelioma. Also, predicted lineage de-differentiation identified neuroendocrine/small cell tumors and was associated with poor outcomes across tumor types. Our platform-independent single-sample approach can be easily translated to existing RNA-seq platforms. ATLAS can complement and guide traditional histopathologic assessment in challenging situations and tumors of unknown primary.

Toxicity and Patient-Reported Quality-of-Life Outcomes After Prostate Stereotactic Body Radiation Therapy With Focal Boost to Magnetic Resonance Imaging-Identified Prostate Cancer Lesions: Results of a Phase 2 Trial

PURPOSE

In this prospective phase 2 trial, we investigated the toxicity and patient-reported quality-of-life outcomes in patients treated with stereotactic body radiation therapy (SBRT) to the prostate gland and a simultaneous focal boost to magnetic resonance imaging (MRI)-identified intraprostatic lesions while also de-escalating dose to the adjacent organs at risk.

METHODS AND MATERIALS

Eligible patients included low- or intermediate-risk prostate cancer (Gleason score ≤7, prostate specific antigen ≤20, T stage ≤2b). SBRT was prescribed to 40 Gy in 5 fractions delivered every other day to the prostate, with any areas of high disease burden (MRI-identified prostate imaging reporting and data system 4 or 5 lesions) simultaneously escalated to 42.5 to 45 Gy and areas overlapping organs at risk (within 2 mm of urethra, rectum, and bladder) constrained to 36.25 Gy (n = 100). Patients without a pretreatment MRI or without MRI-identified lesions were treated to dose of 37.5 Gy with no focal boost (n = 14).

RESULTS

From 2015 to 2022, a total of 114 patients were enrolled with a median follow-up of 42 months. No acute or late grade 3+ gastrointestinal (GI) toxicity was observed. One patient developed late grade 3 genitourinary (GU) toxicity at 16 months. In patients treated with focal boost (n = 100), acute grade 2 GU and GI toxicity was seen in 38% and 4% of patients, respectively. Cumulative late grade 2+ GU and GI toxicities at 24 months were 13% and 5% respectively. Patient-reported outcomes showed no significant long-term change from baseline in urinary, bowel, hormonal, or sexual quality-of-life scores after treatment.

CONCLUSIONS

SBRT to a dose of 40 Gy to the prostate gland with a simultaneous focal boost up to 45 Gy is well tolerated with similar rates of acute and late grade 2+ GI and GU toxicity as seen in other SBRT regimens without intraprostatic boost. Moreover, no significant long-term changes were seen in patient-reported urinary, bowel, or sexual outcomes from pretreatment baseline.

Disease and Toxicity Outcomes after Salvage Radiotherapy (SRT) for Biochemically Recurrent Prostate Cancer (PC) for Patients Enrolled in a Phase II, Open Label Trial Investigating Neoadjuvant Chemohormonal Therapy Followed by Radical Prostatectomy (RP)

PURPOSE/OBJECTIVE(S)

Men with metastatic hormone sensitive prostate cancer have improved progression free and overall survival (PFS/OS) when treated with docetaxel and androgen deprivation therapy (ADT). Our institution conducted a phase II trial (UW17009) investigating the addition of three cycles of neoadjuvant docetaxel and ADT to RP in men with high-risk prostate cancer. Here we report toxicity and freedom from biochemical recurrence (FFBCR) for patients treated with SRT following participation in this trial.

MATERIALS/METHODS

Between January 17, 2018 and August 10, 2021, 28 patients enrolled on UW17009. Patients who had a recurrence and received SRT were identified. Toxicity was assessed using modified LENT (Late Effects of Normal Tissues)/RTOG (Radiation Therapy Oncology Group) criteria, and time to recurrence following SRT was determined for each patient. Additionally, the months of ADT received by each patient was recorded. FFBCR was then determined for this cohort using the method of Kaplan and Meier.

RESULTS

Of 28 patients enrolled on UW17009, 20 (71%) had BCR after RP. Of these, 19 received SRT, representing 68% of the patients enrolled on the trial. The rates of acute grade 1 and grade 2 GU toxicity with SRT were 37% (7) and 53% (10), respectively. The rates of acute grade 1 and grade 2 GI toxicity with SRT were 32% (6) and 37% (7), respectively. On patient experienced a grade 4 genitourinary toxicity during SRT. There were no acute grade 3 or grade 5 toxicities. The rates of late grade 1 and grade 2 GU toxicity were 16% (3) and 11% (2), respectively. The rates of late grade 1 and grade 2 GI toxicity were 11% (2) and 5% (1), respectively. There was 1 (5%) late grade 3 GU toxicity. There were no late grade 3-5 GI toxicities, or late grade 4-5 GU toxicities. The average duration of ADT after prostatectomy was 15 months (range: 0-37 months). According to the method of Kaplan and Meier, mean FFBCR after SRT was 34.0 months at a median of 37.8 months. Median FFBCR had not yet been reached. At the time of analysis, 13 patients remained free from biochemical recurrence after SRT.

CONCLUSION

Rates of biochemical recurrence and SRT following neoadjuvant docetaxel and ADT and then RP were consistent with historical data, as was the rate of successful SRT. Toxicities were also consistent with historical data, though there was notably one acute grade 4 GU toxicity.

Factors Associated with Grade 3-4 Late Toxicity after Hypofractionated Prostate Salvage Radiation

PURPOSE/OBJECTIVE(S)

We previously reported disease outcomes on patients treated with moderately hypofractionated salvage RT (65-70 Gy in 26-28 fractions using imaging guided IMRT) with >13 years of follow-up. The rate of all late surgical/radiation/disease related grade 3-5 toxicities was 27%, which were documented at a median of 8.7 years after the end of radiation. Here we performed an analysis to attempt to identify factors associated with these toxicities.

MATERIALS/METHODS

A total of 161 patients were analyzed; 44 patients experienced 58 late grade ≥3 toxicities. These were analyzed in several groups which included: all (n = 44), GU (n = 40), GU excluding incontinence (n = 35), stricture/fistula related (n = 28), hematuria related (n = 7), and grade 4 stricture/fistula related (n = 8). We investigated patient factors (age, preRT IPSS), surgical factors (clip volume, approach, margin status), and disease factors (stage, Gleason group, PSA nadir, preRT PSA, surgery to RT time). Surgical clip volume was contoured on CT and recorded in cc. Significance was determined using Mann-Whitney U test for continuous variables and Fisher's Exact test for binary variables.

RESULTS

A higher volume of surgical clips in the prostate fossa was found to be significantly related to eventual grade ≥3 stricture/fistula related event (p = 0.05). The mean surgical clip volume was found to be 2.30 cc in those with a documented grade ≥3 stricture/fistula compared to 1.23 cc in those without. Patients with a positive margin had a 30.0% rate of grade ≥3 GU toxicity compared to 16.2% in those with a negative margin (p = 0.03). Surgical clip volume was not found to be significantly related to pathologic stage, nor to eventual biochemical failure (p = 0.799/0.897). A positive margin was associated with a lower rate of biochemical failure after salvage (p = 0.04). Grade 3 events were documented at a median of 7.7 years and grade 4 events at 12.0 years after the end of radiation.

CONCLUSION

Our previous study found a high rate of grade ≥3 toxicities at time points for which there is a paucity of data both in conventional and hypofractionated regimens, particularly in the era of modern surgical and radiation techniques. We also found late toxicities can occur with increasing severity for many years after salvage radiotherapy. This analysis suggests that margin positivity and volume of surgical clips might identify patients at higher risk for late grade ≥3 toxicities, although the etiologies of these toxicities, whether surgery or radiation-related, are uncertain.

Evaluation of kV Triggered Fiducial Tracking for Prostate SBRT Intrafraction Motion Management

PURPOSE/OBJECTIVE(S)

Stereotactic body radiation therapy (SBRT) is a safe and effective treatment for localized prostate cancer. Given the high dose per fraction, a high degree of precision is needed during SBRT treatment delivery. Fiducial marker tracking is a means of monitoring intrafraction motion, ensuring precise treatment delivery and potentially allowing for decreased planning target volume (PTV) margins. We present a 5-year single institution experience to characterize the performance of fiducial tracking.

MATERIALS/METHODS

We analyzed 132 consecutive patients treated at our institution with prostate SBRT between July 2017 and November 2022 using fiducial tracking with the TrueBeam Auto Beam Hold feature. Patients were treated using a 5-fraction dose escalation protocol, with many receiving a simultaneous integrated boost to the GTV up to 45 Gy. All patients underwent pre-treatment cone-beam CT and planar kV imaging with rigid registration to the fiducial markers. Cine kV images were acquired every 3-5 seconds during treatment delivery and the beam was paused when a fiducial deviated from a given 4 mm diameter tolerance region. Additional intrafraction 2D or 3D imaging techniques were then used to realign the fiducials. Data was retrieved from the record and verify system including: number and type of intrafraction images, number of treatment interruptions, number of and magnitude of shifts, and total treatment time.

RESULTS

The average number of cine kV images acquired per fraction and for the total 5 fraction SBRT course were 86.6 ± 26.8 and 425.6 ± 136.8 respectively. Assuming 1.3 mGy/kV image, the fiducial tracking process delivered an average additional 55.3 cGy over the full treatment course. On average, the beam was paused 5.0 ± 2.5 times per fraction (range 1-21 interruptions) and shifts were applied 4.1 ± 2.1 times (range 0-10.2 times) per fraction. Over the 2560 total shifts applied for all the patients, the average(max) translations were 0.6 (12) mm vrt, 0.7 (22) mm lng, and 0.9 (5) mm lat and the average(max) rotations were 0.19 (5.9)° pitch, 0.01 (2.4)° roll, and 0.16 (6.4)° rtn. Overall, 2.4% of shifts were greater than the 4 mm PTV margin and 4.2% of vertical shifts were greater than the 2.5 mm posterior margin. Average fraction treatment time was 27 ± 9.7 minutes.

CONCLUSION

Fiducial tracking affords the potential to increase daily treatment precision when delivering prostate SBRT. Our data indicate that the beam is typically paused several times during a prostate SBRT fraction, with occasional large shifts required, suggesting the importance of intrafraction motion management. Downsides of kV triggered fiducial tracking include extension of treatment times and delivery of additional radiation dose to the patient. We are investigating if this method of motion management might allow for a reduction in PTV margins to reduce normal tissue dosing.

Using 18F-DCFPyL Prostate-Specific Membrane Antigen-Directed Positron Emission Tomography/Magnetic Resonance Imaging to Define Intraprostatic Boosts for Prostate Stereotactic Body Radiation Therapy

Purpose

The recently reported FLAME trial demonstrated a biochemical disease-free survival benefit to using a focal intraprostatic boost to multiparametric magnetic resonance imaging (mpMRI)-identified lesions in men with localized prostate cancer treated with definitive radiation therapy. Prostate-specific membrane antigen (PSMA)-directed positron emission tomography (PET) may identify additional areas of disease. In this work, we investigated using both PSMA PET and mpMRI in planning focal intraprostatic boosts using stereotactic body radiation therapy (SBRT).

Methods and Materials

We evaluated a cohort of patients (n = 13) with localized prostate cancer who were imaged with 2-(3-(1-carboxy-5-[(6-[18F]fluoro-pyridine-2-carbonyl)-amino]-pentyl)-ureido)-pentanedioic acid (¹⁸F-DCFPyL) PET/MRI on a prospective imaging trial before undergoing definitive therapy. The number of lesions concordant (overlapping) and discordant (no overlap) on PET and MRI was assessed. Overlap between concordant lesions was evaluated using the Dice and Jaccard similarity coefficients. Prostate SBRT plans were created fusing the PET/MRI imaging to computed tomography scans acquired the same day. Plans were created using only MRI-identified lesions, only PET-identified lesions, and the combined PET/MRI lesions. Coverage of the intraprostatic lesions and doses to the rectum and urethra were assessed for each of these plans.

Results

The majority of lesions (21/39, 53.8%) were discordant between MRI and PET, with more lesions seen by PET alone (12) than MRI alone (9). Of lesions that were concordant between PET and MRI, there were still areas that did not overlap between scans (average Dice coefficient, 0.34). Prostate SBRT planning using all lesions to define a focal intraprostatic boost provided the best coverage of all lesions without compromising constraints on the rectum and urethra.

Conclusions

Using both mpMRI and PSMA-directed PET may better identify all areas of gross disease within the prostate. Using both imaging modalities could improve the planning of focal intraprostatic boosts.

Standardized Uptake Value for 18F-Fluorodeoxyglucose Is a Marker of Inflammatory State and Immune Infiltrate in Cervical Cancer

PURPOSE

Chemoradiotherapy for locally advanced cervical cancer fails in over a third of patients. Biomarkers with therapeutic implications are therefore needed. We investigated the relationship between an established prognostic marker, maximum standardized uptake value (SUV_max) on ¹⁸F-fluorodeoxyglucose positron emission tomography, and the inflammatory and immune state of cervical cancers.

EXPERIMENTAL DESIGN

An SUV_max most prognostic for freedom from progression (FFP) was identified and compared with known prognostic clinical variables in a cohort of 318 patients treated with definitive radiation with prospectively collected clinical data. Gene set enrichment analysis (GSEA) and CIBERSORT of whole-transcriptome data from 68 patients were used to identify biological pathways and immune cell subpopulations associated with high SUV_max. IHC using a tissue microarray (TMA, N = 82) was used to validate the CIBERSORT findings. The impact of macrophages on cervical cancer glucose metabolism was investigated in coculture experiments.

RESULTS

SUV_max <11.4 was most prognostic for FFP (P = 0.001). The GSEA showed that high SUV_max is associated with increased gene expression of inflammatory pathways, including JAK/STAT3 signaling. CIBERSORT and CD68 staining of the TMA showed high SUV_max tumors are characterized by a monocyte-predominant immune infiltrate. Coculture of cervical cancer cells with macrophages or macrophage-conditioned media altered glucose uptake, and IL6 and JAK/STAT3 signaling contribute to this effect.

CONCLUSIONS

SUV_max is a prognostic marker in cervical cancer that is associated with activation of inflammatory pathways and tumor infiltration of myeloid-derived immune cells, particularly macrophages. Macrophages contribute to changes in cervical cancer glucose metabolism.See related commentary by Williamson et al., p. 4136.

Alteration of Cellular Reduction Potential Will Change 64Cu-ATSM Signal With or Without Hypoxia

Therapies targeting reductive/oxidative (redox) metabolism hold potential in cancers resistant to chemotherapy and radiation. A redox imaging marker would help identify cancers susceptible to redox-directed therapies. Copper(II)-diacetyl-bis(4-methylthiosemicarbazonato) (Cu-ATSM) is a PET tracer developed for hypoxia imaging that could potentially be used for this purpose. We aimed to demonstrate that Cu-ATSM signal is dependent on cellular redox state, irrespective of hypoxia. Methods: We investigated the relationship between ⁶⁴Cu-ATSM signal and redox state in human cervical and colon cancer cells. We altered redox state using drug strategies and single-gene mutations in isocitrate dehydrogenases (IDH1/2). Concentrations of reducing molecules were determined by spectrophotometry and liquid chromatography-mass spectrometry and compared with ⁶⁴Cu-ATSM signal in vitro. Mouse models of cervical cancer were used to evaluate the relationship between ⁶⁴Cu-ATSM signal and levels of reducing molecules in vivo, as well as to evaluate the change in ⁶⁴Cu-ATSM signal after redox-active drug treatment. Results: A correlation exists between baseline ⁶⁴Cu-ATSM signal and cellular concentration of glutathione, nicotinamide adenine dinucleotide phosphate (NADPH), and nicotinamide adenine dinucleotide (NADH). Altering NADH and NADPH metabolism using drug strategies and IDH1 mutations resulted in significant changes in ⁶⁴Cu-ATSM signal under normoxic conditions. Hypoxia likewise changed ⁶⁴Cu-ATSM signal, but treatment of hypoxic cells with redox-active drugs resulted in a more dramatic change than hypoxia alone. A significant difference in NADPH was seen between cervical tumor orthotopic implants in vivo, without a corresponding difference in ⁶⁴Cu-ATSM signal. After treatment with β-lapachone, there was a change in ⁶⁴Cu-ATSM signal in xenograft tumors smaller than 50 mg but not in larger tumors. Conclusion: ⁶⁴Cu-ATSM signal reflects redox state, and altering redox state impacts ⁶⁴Cu-ATSM metabolism. Our animal data suggest there are other modulating factors in vivo. These findings have implications for the use of ⁶⁴Cu-ATSM as a predictive marker for redox therapies, though further in vivo work is needed.

Manipulation of Glucose and Hydroperoxide Metabolism to Improve Radiation Response

Dysregulated glucose and redox metabolism are near universal features of cancers. They therefore represent potential selectively toxic metabolic targets. This review outlines the preclinical and clinical data for targeting glucose and hydroperoxide metabolism in cancer, with a focus on drug strategies that have the most available evidence. In particular, inhibition of glycolysis using 2-deoxyglucose, and inhibition of redox metabolism using the glutathione pathway inhibitor buthionine sulfoximine and the thioredoxin pathway inhibitor auranofin, have shown promise in preclinical studies to increase sensitivity to chemotherapy and radiation by increasing intracellular oxidative stress. Combined inhibition of glycolysis, glutathione, and thioredoxin pathways sensitizes highly glycolytic, radioresistant cancer models in vitro and in vivo. Although the preclinical data support this approach, clinical data are limited to exploratory trials using a single drug in combination with either chemotherapy or radiation. Open research questions include optimizing drug strategies for targeting glycolysis and redox metabolism, determining the appropriate timing for administering this therapy with concurrent chemotherapy and radiation, and identifying biomarkers to determine the cancers that would benefit most from this approach. Given the quality of preclinical evidence, dual targeting of glycolysis and redox metabolism in combination with chemotherapy and radiation should be further evaluated in clinical trials.

Spatial relationship of 2-deoxy-2-[18F]-fluoro-D-glucose positron emission tomography and magnetic resonance diffusion imaging metrics in cervical cancer

Background

This study investigated the spatial relationship of 2-deoxy-2-[¹⁸F]-fluoro-D-glucose positron emission tomography ([¹⁸F]FDG-PET) standardized uptake values (SUVs) and apparent diffusion coefficients (ADCs) derived from magnetic resonance (MR) diffusion imaging on a voxel level using simultaneously acquired PET/MR data.

We performed an institutional retrospective analysis of patients with newly diagnosed cervical cancer who received a pre-treatment simultaneously acquired [¹⁸F]FDG-PET/MR. Voxel SUV and ADC values, and global tumor metrics including maximum SUV (SUV_max), mean ADC (ADC_mean), and mean tumor-to-muscle ADC ratio (ADC_T/M) were compared. The impacts of histology, grade, and tumor volume on the voxel SUV to ADC relationship were also evaluated. The potential prognostic value of the voxel SUV/ADC relationship was evaluated in an exploratory analysis using Kaplan-Meier/log-rank and univariate Cox analysis.

Results

Seventeen patients with PET/MR scans were identified. There was a significant inverse correlation between SUV_max and ADC_mean, and SUV_max and ADC_T/M. In the voxelwise analysis, squamous cell carcinomas (SCCAs) and poorly differentiated tumors showed a consistent significant inverse correlation between voxel SUV and ADC values; adenocarcinomas (AdenoCAs) and well/moderately differentiated tumors did not. The strength of the voxel SUV/ADC correlation varied with metabolic tumor volume (MTV). On log-rank analysis, the correlation between voxel SUV/ADC values was prognostic of disease-free survival (DFS).

Conclusions

In this hypothesis-generating study, a consistent inverse correlation between voxel SUV and ADC values was seen in SCCAs and poorly differentiated tumors. On univariate statistical analysis, correlation between voxel SUV and ADC values was prognostic for DFS.

Electronic supplementary material

The online version of this article (10.1186/s13550-018-0403-7) contains supplementary material, which is available to authorized users.

Radioresistant Cervical Cancers Are Sensitive to Inhibition of Glycolysis and Redox Metabolism

Highly glycolytic cervical cancers largely resist treatment by cisplatin and coadministered pelvic irradiation as the present standard of care. In this study, we investigated the effects of inhibiting glycolysis and thiol redox metabolism to evaluate them as alternate treatment strategies in these cancers. In a panel of multiple cervical cancer cell lines, we evaluated sensitivity to inhibition of glycolysis (2-deoxyglucose, 2-DG) with or without simultaneous inhibition of glutathione and thioredoxin metabolism (BSO/AUR). Intracellular levels of total and oxidized glutathione, thioredoxin reductase activity, and indirect measures of intracellular reactive oxygen species were compared before and after treatment. Highly radioresistant cells were the most sensitive to 2-DG, whereas intermediate radioresistant cells were sensitive to 2-DG plus BSO/AUR. In response to 2-DG/BSO/AUR treatment, we observed increased levels of intracellular oxidized glutathione, redox-sensitive dye oxidation, and decreased glucose utilization via multiple metabolic pathways including the tricarboxylic acid cycle. 2-DG/BSO/AUR treatment delayed the growth of tumors composed of intermediate radioresistant cells and effectively radiosensitized these tumors at clinically relevant radiation doses both in vitro and in vivo Overall, our results support inhibition of glycolysis and intracellular redox metabolism as an effective alternative drug strategy for the treatment of highly glycolytic and radioresistant cervical cancers.Significance: This study suggests a simple metabolic approach to strike at an apparent Achilles' heel in highly glycolytic, radioresistant forms of cervical cancers, possibly with broader applications in cancer therapy. Cancer Res; 78(6); 1392-403. ©2018 AACR.

The sensitivity and specificity of F-DOPA PET in a movement disorder clinic

Idiopathic Parkinson's disease (PD) is the second most common neurodegenerative disorder. Early PD may present a diagnostic challenge with broad differential diagnoses that are not associated with nigral degeneration or striatal dopamine deficiency. Therefore, the early clinical diagnosis alone may not be accurate and this reinforces the importance of functional imaging targeting the pathophysiology of the disease process. (18)F-DOPA L-6-[(18)F] fluoro-3,4-dihydroxyphenylalnine ((18)F-DOPA) is a positron emission tomography (PET) agent that measures the uptake of dopamine precursors for assessment of presynaptic dopaminergic integrity and has been shown to accurately reflect the monoaminergic disturbances in PD. In this study, we aim to illustrate our local experience to determine the accuracy of (18)F-DOPA PET for diagnosis of PD. We studied a total of 27 patients. A retrospective analysis was carried out for all patients that underwent (18)F-DOPA PET brain scan for motor symptoms suspicious for PD between 2001-2008. Both qualitative and semi-quantitative analyses of the scans were performed. The patient's medical records were then assessed for length of follow-up, response to levodopa, clinical course of illness, and laterality of symptoms at time of (18)F-DOPA PET. The eventual diagnosis by the referring neurologist, movement disorder specialist, was used as the reference standard for further analysis. Of the 28 scans, we found that one was a false negative, 20 were true positives, and 7 were true negatives. The resultant values are Sensitivity 95.4% (95% CI: 100%-75.3%), Specificity 100% (95% CI: 100%-59.0%), PPV 100% (95% CI 100%-80.7%), and NPV 87.5% (95% CI: 99.5%-50.5%).

Alkylphosphocholine analogs for broad-spectrum cancer imaging and therapy

Many solid tumors contain an overabundance of phospholipid ethers relative to normal cells. Capitalizing on this difference, we created cancer-targeted alkylphosphocholine (APC) analogs through structure-activity analyses. Depending on the iodine isotope used, radioiodinated APC analog CLR1404 was used as either a positron emission tomography (PET) imaging ((124)I) or molecular radiotherapeutic ((131)I) agent. CLR1404 analogs displayed prolonged tumor-selective retention in 55 in vivo rodent and human cancer and cancer stem cell models. (131)I-CLR1404 also displayed efficacy (tumor growth suppression and survival extension) in a wide range of human tumor xenograft models. Human PET/CT (computed tomography) and SPECT (single-photon emission computed tomography)/CT imaging in advanced-cancer patients with (124)I-CLR1404 or (131)I-CLR1404, respectively, demonstrated selective uptake and prolonged retention in both primary and metastatic malignant tumors. Combined application of these chemically identical APC-based radioisosteres will enable personalized dual modality cancer therapy of using molecular (124)I-CLR1404 tumor imaging for planning (131)I-CLR1404 therapy.

Nonlinear spatio-temporal filtering of dynamic PET data using a four-dimensional Gaussian filter and expectation-maximization deconvolution

We introduce a method for denoising dynamic PET data, spatio-temporal expectation-maximization (STEM) filtering, that combines four-dimensional Gaussian filtering withEMdeconvolution. The initial Gaussian filter suppresses noise at a broad range of spatial and temporal frequencies and EM deconvolution quickly restores the frequencies most important to the signal. We aim to demonstrate that STEM filtering can improve variance in both individual time frames and in parametric images without introducing significant bias. We evaluate STEM filtering with a dynamic phantom study, and with simulated and human dynamic PET studies of a tracer with reversible binding behaviour, [C-11]raclopride, and a tracer with irreversible binding behaviour, [F-18]FDOPA. STEM filtering is compared to a number of established three and four-dimensional denoising methods. STEM filtering provides substantial improvements in variance in both individual time frames and in parametric images generated with a number of kinetic analysis techniques while introducing little bias. STEM filtering does bias early frames, but this does not affect quantitative parameter estimates. STEM filtering is shown to be superior to the other simple denoising methods studied. STEM filtering is a simple and effective denoising method that could be valuable for a wide range of dynamic PET applications.

(18)F-DOPA PET with and without MRI fusion, a receiver operator characteristics comparison

This study is a retrospective analysis of the diagnostic accuracy of FDOPA PET with MRI fusion to FDOPA PET without MRI fusion. Clinical FDOPA PET scans obtained between 2000 and 2008 at the University of Wisconsin Hospital and Clinics were assessed using measures derived from regions of interest (ROI) generated with fused MRI (fused group) and again with ROIs derived solely from PET data (non-fused groups). The ROIs were used to calculate ratios (Striatum/Occipital cortex, Striatum/Cerebellum) pertinent to Parkinson's disease (PD) pathology. The clinical records were assessed for demographic data, follow-up length, and diagnosis. Receiver Operator Characteristics with area under the curve (AUC) measures were calculated and compared using confidence intervals and hypothesis testing. 27 patients had FDOPA PET with median clinical follow-up of 4 years. Of these, 17 patients had FDOPA PET with a fusible MR image. Seven of the 27 had a non-PD movement disorder. AUCs for the ratio measures ranged from 0.97-1.0 (fused), 0.73-0.83 (non-fused), and 0.63-0.82 (matched non-fused). The fused images had improved accuracy compared to the matched non-fused and all non-fused groups for the striatum to occipital group (p=0.04, p=0.03), while the striatum to cerebellum ratio had improvement over the non-fused all group (p=0.041). MR fusion to FDOPA PET improves the accuracy of at least some measures (Striatum/Occiput, Striatum/Cerebellum) in the diagnosis of PD.

Improved kinetic analysis of dynamic PET data with optimized HYPR-LR

PURPOSE

Highly constrained backprojection-local reconstruction (HYPR-LR) has made a dramatic impact on magnetic resonance angiography (MRA) and shows promise for positron emission tomography (PET) because of the improvements in the signal-to-noise ratio (SNR) it provides dynamic images. For PET in particular, HYPR-LR could improve kinetic analysis methods that are sensitive to noise. In this work, the authors closely examine the performance of HYPR-LR in the context of kinetic analysis, they develop an implementation of the algorithm that can be tailored to specific PET imaging tasks to minimize bias and maximize improvement in variance, and they provide a framework for validating the use of HYPR-LR processing for a particular imaging task.

METHODS

HYPR-LR can introduce errors into non sparse PET studies that might bias kinetic parameter estimates. An implementation of HYPR-LR is proposed that uses multiple temporally summed composite images that are formed based on the kinetics of the tracer being studied (HYPR-LR-MC). The effects of HYPR-LR-MC and of HYPR-LR using a full composite formed with all the frames in the study (HYPR-LR-FC) on the kinetic analysis of Pittsburgh compound-B ([11C]-PIB) are studied. HYPR-LR processing is compared to spatial smoothing. HYPR-LR processing was evaluated using both simulated and human studies. Nondisplaceable binding potential (BP(ND)) parametric images were generated from fifty noise realizations of the same numerical phantom and eight [(11)C]-PIB positive human scans before and after HYPR-LR processing or smoothing using the reference region Logan graphical method and receptor parametric mapping (RPM2). The bias and coefficient of variation in the frontal and parietal cortex in the simulated parametric images were calculated to evaluate the absolute performance of HYPR-LR processing. Bias in the human data was evaluated by comparing parametric image BP(ND) values averaged over large regions of interest (ROIs) to Logan estimates of the BP(ND) from TACs averaged over the same ROIs. Variance was assessed qualitatively in the parametric images and semiquantitatively by studying the correlation between voxel BP(ND) estimates from Logan analysis and RPM2.

RESULTS

Both the simulated and human data show that HYPR-LR-FC overestimates BP(ND) values in regions of high [(11)C]-PIB uptake. HYPR-LR-MC virtually eliminates this bias. Both implementations of HYPR-LR reduce variance in the parametric images generated with both Logan analysis and RPM2, and HYPR-LR-FC provides a greater reduction in variance. This reduction in variance nearly eliminates the noise-dependent Logan bias. The variance reduction is greater for the Logan method, particularly for HYPR-LR-MC, and the variance in the resulting Logan images is comparable to that in the RPM2 images. HYPR-LR processing compares favorably with spatial smoothing, particularly when the data are analyzed with the Logan method, as it provides a reduction in variance with no loss of spatial resolution.

CONCLUSIONS

HYPR-LR processing shows significant potential for reducing variance in parametric images, and can eliminate the noise-dependent Logan bias. HYPR-LR-FC processing provides the greatest reduction in variance but introduces a positive bias into the BP(ND) of high-uptake border regions. The proposed method for forming HYPR composite images, HYPR-LR-MC, eliminates this bias at the cost of less variance reduction.

Impact of expectation-maximization reconstruction iterations on the diagnosis of temporal lobe epilepsy with PET

There is a well known tradeoff between image noise and image sharpness that is dependent on the number of iterations performed in ordered subset expectation maximization (OSEM) reconstruction of PET data. We aim to evaluate the impact of this tradeoff on the sensitivity and specificity of (18)F-FDG PET for the diagnosis of temporal lobe epilepsy. A retrospective blinded reader study was performed on two OSEM reconstructions, using either 2 or 5 iterations, of 32 (18)F-FDG PET studies acquired at our institution for the diagnosis of temporal lobe epilepsy. The sensitivity and specificity of each reconstruction for identifying patients who were ultimately determined to be surgical candidates was assessed using an ROC analysis. The sensitivity of each reconstruction for identifying patients who showed clinical improvement following surgery was also assessed. Our results showed no significant difference between the two reconstructions studied for either the sensitivity and specificity of (18)F-FDG PET for predicting surgical candidacy, or its sensitivity for predicting positive surgical outcomes. This implies that the number of iterations performed during OSEM reconstruction will have little impact on a reader based interpretation of (18)F-FDG PET scans acquired for the diagnosis of temporal lobe epilepsy, and can be determined by physician and institutional preference.

Application of a whole-body pharmacokinetic model for targeted radionuclide therapy to NM404 and FLT

We have previously developed a model that provides relative dosimetry estimates for targeted radionuclide therapy (TRT) agents. The whole-body and tumor pharmacokinetic (PK) parameters of this model can be noninvasively measured with molecular imaging, providing a means of comparing potential TRT agents. Parameter sensitivities and noise will affect the accuracy and precision of the estimated PK values and hence dosimetry estimates. The aim of this work is to apply a PK model for TRT to two agents with different magnitudes of clearance rates, NM404 and FLT, explore parameter sensitivity with respect to time and investigate the effect of noise on parameter precision and accuracy. Twenty-three tumor bearing mice were injected with a 'slow-clearing' agent, (124)I-NM404 (n = 10), or a 'fast-clearing' agent, (18)F-FLT (3'-deoxy-3'-fluorothymidine) (n = 13) and imaged via micro-PET/CT pseudo-dynamically or dynamically, respectively. Regions of interest were drawn within the heart and tumor to create time-concentration curves for blood pool and tumor. PK analysis was performed to estimate the mean and standard error of the central compartment efflux-to-influx ratio (k(12)/k(21)), central elimination rate constant (k(el)), and tumor influx-to-efflux ratio (k(34)/k(43)), as well as the mean and standard deviation of the dosimetry estimates. NM404 and FLT parameter estimation results were used to analyze model accuracy and parameter sensitivity. The accuracy of the experimental sampling schedule was compared to that of an optimal sampling schedule found using Cramer-Rao lower bounds theory. Accuracy was assessed using correlation coefficient, bias and standard error of the estimate normalized to the mean (SEE/mean). The PK parameter estimation of NM404 yielded a central clearance, k(el) (0.009 ± 0.003 h(-1)), normal body retention, k(12)/k(21) (0.69 ± 0.16), tumor retention, k(34)/k(43) (1.44 ± 0.46) and predicted dosimetry, D(tumor) (3.47 ± 1.24 Gy). The PK parameter estimation of FLT yielded a central elimination rate constant, k(el) (0.050 ± 0.025 min(-1)), normal body retention, k(12)/k(21) (2.21 ± 0.62) and tumor retention, k(34)/k(43) (0.65 ± 0.17), and predicted dosimetry, D(tumor) (0.61 ± 0.20 Gy). Compared to experimental sampling, optimal sampling decreases the dosimetry bias and SEE/mean for NM404; however, it increases bias and decreases SEE/mean for FLT. For both NM404 and FLT, central compartment efflux rate constant, k(12), and central compartment influx rate constant, k(21), possess mirroring sensitivities at relatively early time points. The instantaneous concentration in the blood, C(0), was most sensitive at early time points; central elimination, k(el), and tumor efflux, k(43), are most sensitive at later time points. A PK model for TRT was applied to both a slow-clearing, NM404, and a fast-clearing, FLT, agents in a xenograft murine model. NM404 possesses more favorable PK values according to the PK TRT model. The precise and accurate measurement of k(12), k(21), k(el), k(34) and k(43) will translate into improved and precise dosimetry estimations. This work will guide the future use of this PK model for assessing the relative effectiveness of potential TRT agents.

Surgical decision making in temporal lobe epilepsy: a comparison of [(18)F]FDG-PET, MRI, and EEG

OBJECTIVES

The goals of this work were (1) to determine the effect of [(18)F]fluorodeoxyglucose positron emission tomography (FDG-PET), MRI, and EEG on the decision to perform temporal lobe epilepsy (TLE) surgery, and (2) to determine if FDG-PET, MRI, or EEG predicts surgical outcome.

METHODS

All PET scans ordered (2000-2010) for epilepsy or seizures were tabulated. Medical records were investigated to determine eligibility and collect data. Statistical analysis included odds ratios, κ statistics, univariate analysis, and logistic regression.

RESULTS

Of the 186 patients who underwent FDG-PET, 124 had TLE, 50 were surgical candidates, and 34 had surgery with post-operative follow-up. Median length of follow-up was 24 months. MRI, FDG-PET, and EEG were significant predictors of surgical candidacy (P<0.001) with odds ratios of 42.8, 20.4, and 6.3, respectively. FDG-PET was the only significant predictor of postoperative outcome (P<0.01).

CONCLUSION

MRI showed a trend toward having the most influence on surgical candidacy, but only FDG-PET predicted surgical outcome.

Dynamic PET denoising with HYPR processing

HighlY constrained backPRojection (HYPR) is a promising image-processing strategy with widespread application in time-resolved MRI that is also well suited for PET applications requiring time series data. The HYPR technique involves the creation of a composite image from the entire time series. The individual time frames then provide the basis for weighting matrices of the composite. The signal-to-noise ratio (SNR) of the individual time frames can be dramatically improved using the high SNR of the composite image. In this study, we introduced the modified HYPR algorithm (the HYPR method constraining the backprojections to local regions of interest [HYPR-LR]) for the processing of dynamic PET studies. We demonstrated the performance of HYPR-LR in phantom, small-animal, and human studies using qualitative, semiquantitative, and quantitative comparisons. The results demonstrate that significant improvements in SNR can be realized in the PET time series, particularly for voxel-based analysis, without sacrificing spatial resolution. HYPR-LR processing holds great potential in nuclear medicine imaging for all applications with low SNR in dynamic scans, including for the generation of voxel-based parametric images and visualization of rapid radiotracer uptake and distribution.