Interim PET Evaluation in Diffuse Large B-Cell Lymphoma Using Published Recommendations: Comparison of the Deauville 5-Point Scale and the ΔSUVmax Method.
Jan Rekowski, Andreas Hüttmann, Christine Schmitz, Stefan P. Müller, Lars Kurch, Jörg Kotzerke, Christiane Franzius, Matthias Weckesser, Frank M. Bengel, Martin Freesmeyer, Andreas Hertel, Thomas Krohn, Jens Holzinger, Ingo Brink, Uwe Haberkorn, Fonyuy Nyuyki, Daniëlle M.E. van Assema, Lilli Geworski, Dirk Hasenclever, Karl-Heinz Jöckel, and Ulrich Dührsen.
J Nucl Med 2021; 62(1): 37-42.
Background: Interim 18F-fluorodeoxyglucose PET/CT is performed among chemotherapy cycles to assess treatment response and guide further management in aggressive lymphomas such as diffuse large B-cell lymphoma (DLBCL). The current recommendation for analysing treatment response in these interim PET is by the Deauville score. This five-point scale assigns a value based on relative radiotracer uptake in the disease sites: 1 is no residual uptake, 2 is uptake less than the mediastinum, 3 is uptake less than or equal to the liver, 4 is uptake somewhat greater than the liver, and 5 is uptake much greater than the liver; an unfavourable treatment response is accepted as a score of 4 or 5. The Deauville score requires only the interim PET/CT, but is associated with greater false-positives and thus underestimation of treatment response. Treatment response could also be assessed by ΔSUVmax, which compares the disease site with highest radiotracer uptake (SUVmax) between the first PET/CT and the interim PET/CT (ΔSUVmax); an unfavourable treatment response is accepted as two-thirds reduction in uptake or less. The ΔSUVmax method requires a pre-treatment and interim PET/CT, but provides a semi-quantitative assessment unaffected by background uptake. There are few studies comparing the two methods, and more work is needed.
Purpose: To compare the Deauville score and ΔSUVmax method for assessing treatment response in DLBCL
Methods: Data from the multicentre PET-guided Therapy of Aggressive non-Hodgkin Lymphomas (PETAL) randomized control trial were used: this trial investigated treatment options in newly diagnosed non-Hodgkin lymphomas (NHL) based on interim PET/CT response. All patients received the same initial chemotherapy regimen. Those with unfavourable interim PET/CT results underwent a different subsequent treatment than those with favourable results: final patient outcomes were unaffected by the treatment change. All patients received baseline and interim PET/CT, with the interim scan performed at least ten days after the second chemotherapy cycle. Interim PET/CT scans were assessed with the ΔSUVmax method: response was deemed favourable if there was greater than two-thirds reduction in SUVmax or if the amount of uptake was physiologic. Retrospectively, these scans were also assessed with the Deauville score with a favourable response being of score of 1 to 3. The authors used Spearmans’ rank correlation coefficient to analyse the association between the two methods, Kaplan-Meier curves to analyse outcome differences, and AUC for discrimination performance.
Results: 596 patients were included. 92 had an unfavourable response by ΔSUVmax (two-thirds or less reduction), 29 of whom had only physiologic levels of uptake on the interim PET/CT and were therefore re-classified as favourable, leaving 63 unfavourable responses by ΔSUVmax. By Deauville score, 270 had an unfavourable response. Concordance between the two methods was 63%, over one third having unfavourable response by Deauville but favourable by ΔSUVmax.
Event-free survival at two years was 71%. Discrimination performance was poor for both methods, but was higher for the ΔSUVmax method than for the Deauville score. Kaplan-Meier event-free survival curves illustrated more pronounced differences in favourable/unfavourable groups by ΔSUVmax than by Deauville score. The Deauville score was found to have a higher sensitivity (52.5% vs 24.6%) but lower specificity (57.5% vs 88.8%) than the ΔSUVmax method, thus a higher false-positive rate. The ΔSUVmax method had a higher positive predictive value and lower negative predictive value than the Deauville score in all outcomes.
Discussion: Previous studies showed that interim PET/CT is predictive of overall survival only when analysed by ΔSUVmax method, and that the Deauville score is associated with a high false-positive rate. In the present study, the ΔSUVmax method was shown to have better discrimination performance than the Deauville score, and a high false-positive rate with the Deauville score was redemonstrated. Assessing interim PET/CT with the Deauville score would therefore be more likely to suggest that a patient requires a more aggressive therapy, though they may indeed have responded well to their current, less toxic and less expensive therapy.
Conclusion: Where baseline and interim PET/CT is available, the ΔSUVmax method may provide better prognostic value in treatment decisions based on early metabolic response given its lower false-positive rate compared to the Deauville score.
Assessment of Bone Lesions with 18F-FDG PET Compared with 99mTc Bone Scintigraphy Leads to Clinically Relevant Differences in Metastatic Breast Cancer Management.
Suzanne C. van Es, Ton Velleman, Sjoerd G. Elias, Frederike Bensch, Adrienne H. Brouwers, Andor W.J.M. Glaudemans, Thomas C. Kwee, Marleen Woltman-van Iersel, John H. Maduro, Sjoukje F. Oosting, Elisabeth G.E. de Vries, and Carolina P. Schröder.
J Nucl Med 2021; 62(2): 177-83.
Background: Bone is the first site of metastasis for half of all patients who develop metastatic breast cancer. Currently, the National Comprehensive Cancer Network (NCCN) recommend whole body bone scan (BS) to assess for the presence of bone metastases and consider 18F-fluorodeoxyglucose PET (FDG-PET) to be an optional additional modality. The use of BS alone can risk missing osteolytic metastases. While there are studies comparing BS and FDG-PET, they do not compare clinical implications.
Purpose: To compare clinical outcomes in patients with metastatic breast cancer (MBC) following BS with contrast-enhanced CT (ceCT) and FDG-PET with ceCT
Methods: Patients with newly diagnosed non-rapidly progressive MBC who had been enrolled in the multicentre IMPACT-MBC were included. All patients underwent BS, FDG-PET/CT, and ceCT at baseline. Bone lesions were assessed on all three modalities by two radiologists and two nuclear medicine specialists. Management recommendations were made by a panel of five oncologists on two separate occasions over ten meetings, based on the BS+ceCT during the first discussion and on the FDG-PET/CT+ceCT during the second. Any relevant information regarding visceral organ involvement provided by the FDG-PET was presented during both discussions. When analysing these recommendations, clinically relevant differences included different treatment intent (curative or not) and different systemic or local therapies.
Results: Of the 102 patients with MBC included, 93 had bone lesions on either ceCT, BS, and/or FDG-PET, with 9 patients having equivocal bone lesions. Among the other 84 patients, 3,473 bone lesions were identified: 1,004 on ceCT alone, 655 on BS alone, and 2,285 on FDG-PET alone. FDG-PET identified significantly more bone lesions than ceCT and BS, and ceCT more than BS. ceCT identified 299 lesions in 33 patients which were not seen on BS or FDG-PET.
Management recommendations differed in 16 of 102 patients as lesions were identified on FDG-PET+ceCT which were not seen on BS+ceCT, resulting in more systemic therapy (7 patients) or treatment intent changing to palliative (9 patients). Those for whom more systemic therapy was recommended had extensive metastatic disease with visceral involvement, while those for whom intent was changed had otherwise limited metastatic disease. In 27 patients, recommendations would have included performing an FDG-PET to further evaluate the lesions.
Discussion: FDG-PET+ceCT resulted in changes to management recommendations in 16 patients (9 changes to palliative intent, 7 changes to more systemic therapy). While no prior studies have compared the clinical relevance of BS and FDG-PET for MBC, studies have compared their ability to detect bone metastases: retrospective studies have been done which suggest that, if FDG-PET is to be done, no BS is required as BS will not show lesions which are not identifiable on FDG-PET. The present work redemonstrated that FDG-PET identifies significantly more lesions than BS, and suggests that the use of FDG-PET+ceCT over BS+ceCT could lead to clinically relevant management changes.
Hybrid imaging in dementia: A semi-quantitative (18F)-fluorodeoxyglucose positron emission tomography/magnetic resonance imaging approach in clinical practice.
Ana Marija Franceschi, Kiyon Naser-Tavakolian, Michael Clifton, Osama Ahmed, Katarina Stoffers, Lev Bangiyev, Giuseppe Cruciata, Sean Clouston, and Dinko Franceschi.
World J Nucl Med 2021; 20(1): 23-31.
Background: PET/MRI is emerging as a new means of evaluating neurodegenerative disorders. Early diagnosis of neurodegenerative disorders is key to effective treatment to maintain quality of life and slow disease progression. 18F-fluorodeoxyglucose PET (FDG-PET) brain imaging illustrates characteristics of the varieties of neurodegenerative disorders, such as specific lobe involvement and patterns of hypometabolism: Alzheimer’s disease (AD), for example, involves the medial temporal and posterior parietal lobes with corresponding hypometabolism; frontotemporal dementia (FTD) typically involves the frontal and temporal lobes; dementia with Lewy bodies (DLB) affects the occipital lobe; finally, corticobasal degeneration (CBD), while less well understood given its rarity, affects the frontal and parietal lobes. Coupling FDG-PET with simultaneous MRI allows for improved anatomic localization and correction of atrophic regions which would underestimate local metabolism.
Purpose: To assess the relationship between grey-matter volume loss and decreased metabolic uptake of specific lobes of the brain in patients undergoing FDG-PET/MRI for clinical dementia
Methods: This retrospective study identified 89 patients who underwent dedicated brain FDG-PET/MRI for clinical dementia. All patients subsequently received a diagnosis of a specific subtype of dementia (AD, FTD, DLB, or CBD) based on clinical data and metabolic uptake and corresponding volume loss on FDG-PET/MRI. The authors collected common presenting symptoms from the clinical data, defined as symptoms present in >50% of patients with that subtype. Imaging was performed with a Siemens mMR 3T PET/MRI scanner, and assessments were performed by two neuroradiologists and one nuclear medicine specialist with MIMneuro and NeuroQuant software. To evaluate hypometabolism, semi-quantitative z-scores below -1.65 were defined as abnormal metabolism supportive of the diagnosis for that subtype. To evaluate lobar volume loss, volumes below 5th percentile for age were defined as abnormal volume loss. Diagnoses were discussed among the three radiologists and the referring clinician. Statistical correlation between hypometabolism and volume loss was done with Pearson’s correlation coefficient for each subtype of dementia.
Results: Of the 89 patients, 29 were diagnosed with AD, 34 with FTD, 14 with DLB, and 12 with CBD. In patients with AD, distinct common clinical symptoms included greater than one-year progressive memory loss and difficulties with activities of daily living. A weakly positive relationship between hypometabolism z-score and volume loss percentiles in the parietal (r=0.3, p=0.120) and temporal (r=0.38, p<0.05) lobes was found. In patients with FTD, behavioural changes were unique among the most common presenting symptom. A weakly positive relationship was found in the frontal lobe (r=0.35, p=0.051), with no relationship identified for the temporal lobe (r = 0.02, p=0.916). In patients with DLB, gait disturbance and visual hallucinations were common symptoms. A moderately positive relationship was found for the occipital lobe (r=0.42, p=0.130), and a weakly positive relationship for the parietal lobe (r=0.22, p=0.447). In patients with CBD, distinct common presenting symptoms included unilateral weakness and apraxia. A moderately positive relationship between hypometabolism z-score and volume loss percentiles was found for the parietal lobe (r=0.58, p<0.05).
Discussion: The authors investigated the feasibility of using semi-quantitative lobar-specific analysis of hypometabolism and volume loss to direct diagnosis of dementia subtypes. They demonstrated positive relationships between them in all subtypes, though of varying strengths and statistical significance within each subtype. In the context of the small sample size, these early results suggest there is utility in these semi-quantitative tools to improve one’s confidence in diagnosing dementia subtypes.
References