Amide proton transfer-weighted MRI in the clinical setting – correlation with dynamic susceptibility contrast perfusion in the post-treatment management of adult glioma patients at 3T

Abstract

Abstract (Dansk)

Baggrund: Konventionelle MR sekvenser har begrænset specificitet, når der skal skelnes mellem sygdomsprogression og behandlingsfølger hos patienter med hjerne tumorer (gliomer). Avancerede MR teknikker såsom dynamisk susceptibilitet kontrast (DSC) perfusion benyttes til at kompensere for konventionel MR’s begrænsninger, men kræver gadolinium kontrast anvendelse. Nye avancerede teknikker (amid proton overførsel-vægtet MR - APTw MR), som ikke forudsætter kontrast anvendelse, er derfor interessante at undersøge. Formål: At undersøge korrelationen mellem APTw MR og DSC perfusion i den billeddiagnostiske opfølgning af patienter med hjerne gliomer; at udføre inter-observer reliabilitetsanalyse på baggrund af APTw data. Studie design: Prospektiv tværsnitsstudie. Population: Fyrre voksne patienter, behandlet for histopatologi-verificerede hjerne gliomer: 3 oligodendrogliomer (WHO II), 9 diffuse astrocytomer (WHO II), 4 anaplastiske astrocytomer (WHO III), 4 anaplastiske oligodendrogliomer (WHO III), og 20 glioblastomer (WHO IV). Feltstyrke/ MR pulssekvens: 3T/ 3D APTw MR og DSC perfusion. Billedanalyse: To overlæger med speciale i neuroradiologi udførte region of interest (ROI) målinger på relativ cerebralt blodvolumen (rCBV) og APTw maps. Kontrast-forstærkede T1-vægtet eller FLAIR billeder blev anvendt som anatomisk underlag. Statistisk analyse: Spearman’s test blev anvendt til at vurdere korrelationen mellem maks. APTw (APTwmax) og maks. rCBV (rCBVmax) værdier, intraclass korrelationskoefficienten (ICC) blev benyttet som reliabilitetsmål, og målingsdifferenserne mellem observatører blev visualiseret vha. Bland-Altman plots. Resultater: Vi fandt en statistisk signifikant moderat positiv korrelation mellem APTwmax og rCBVmax (observatør 1: r=0.73; p<0.0001; observatør 2: r=0.62; p<0.0001). Ligeledes var der god inter-observer reliabilitet for APTwmax ifølge ICC (ICC=0.82; 95 % konfidensinterval (CI) 0.66-0.90). De yderste grænser af 95 % konfidensintervallene for Bland-Altman Limits of Agreement var dog lidt brede (-1.60 to 1.90), hvilket tyder på en vis observatør afhængighed af APTw målingerne. ICC for APTwmax var højere end ICC for rCBVmax (ICC=0.74; 95 % CI 0.50-0.86). Alligevel kunne forskellen mellem ICC for APTwmax og rCBVmax ikke betragtes som statistisk signifikant, da 95 % konfidensintervallene overlappede. Konklusion: Der findes en positiv korrelation mellem APTwmax og rCBVmax ved patienter behandlet for hjerne gliomer. APTw teknikken har god reliabilitet og en vis observatør-afhængighed.

Abstract Background: Post-treatment glioma imaging is challenging because of the limited specificity of conventional MRI. Advanced imaging techniques are needed to overcome these limitations, preferably without the use of gadolinium contrast agents. Purpose: To investigate the correlation between amide proton transfer-weighted (APTw) MRI and dynamic susceptibility contrast (DSC) perfusion in glioma follow-up imaging; to perform inter-observer agreement and reliability analysis on APTw data. Study type: Prospective/ cross-sectional. Population: Forty adult subjects, treated for histopathologically confirmed glioma: 3 oligodendroglioma (WHO II), 9 diffuse astrocytoma (WHO II), 4 anaplastic astrocytoma (WHO III), 4 anaplastic oligodendroglioma (WHO III) and 20 glioblastoma (WHO IV). Field strength/ Sequence: 3T/ 3D APTw and DSC perfusion. Assessment: Region of interest (ROI) measurements performed independently by 2 consultant neuroradiologists on relative cerebral blood volume (rCBV) and APTw maps, overlaid on anatomical images. Statistical tests: Spearman’s rank-order was used to test the correlation between maximum APTw (APTwmax) and maximum rCBV (rCBVmax), the intraclass correlation coefficient (ICC) to assess APTw reliability and Bland-Altman (BA) plots to visualize inter-observer agreement. Results: A statistically significant moderately strong positive correlation was observed between APTwmax and rCBVmax (observer 1: r=0.73; p<0.0001; observer 2: r=0.62; p<0.0001). We found good inter-observer agreement for APTwmax (ICC=0.82; 95% confidence interval (CI) 0.66-0.90), although with somewhat broad outer 95% CI of Limits of Agreement (LoA) on the BA plots (-1.60 to 1.90). This indicates some observer dependence of APTw measurements. ICC for APTwmax was higher than ICC for rCBVmax (ICC=0.74; 95% CI 0.50-0.86), but the difference didn’t attain statistical significance because of overlapping 95% confidence intervals (CI). Conclusion: APTwmax values correlate positively with rCBVmax in post-treatment gliomas. APTw imaging is a reliable technique with some observer dependence.