Inter-observer variability of 90Y PET/CT dosimetry in hepatocellular carcinoma after glass microspheres transarterial radioembolization

Abstract

Introduction Strong correlation has been demonstrated between tumor dose and response and between healthy liver dose and side effects. Individualized dosimetry is increasingly recommended in the current clinical routine. However, hepatic and tumor segmentations could be complex in some cases. The aim of this study is to assess the reproducibility of the tumoral and non-tumoral liver dosimetry in selective internal radiation therapy (SIRT). Material and methods Twenty-three patients with hepatocellular carcinoma (HCC) who underwent SIRT with glass microspheres were retrospectively included in the study. Tumor (TV) and total liver volumes (TLV), and mean absorbed doses in tumoral liver (TD) and non-tumoral liver (THLD) were determined on the 90Y PET/CT studies using Simplicit90YTM software, by three independent observers. Dosimetry datasets were obtained by a medical physicist helped by a nuclear medicine (NM) physician with 10 years of experience (A), by a NM physician with 4-year experience (B), and by a resident who first performed 10 dosimetry assessments as a training (C). Inter-observer agreement was evaluated using intra-class correlation coefficients (ICC), coefficients of variation (CV), Bland-Altman plots, and reproducibility coefficient (RDC). Results A strong agreement was observed between all three readers for estimating TLV (ICC 0.98) and THLD (ICC 0.97). Agreement was lower for TV delineation (ICC 0.94) and particularly for TD (ICC 0.73), especially for the highest values. Regarding TD, the CV (%) was 26.5, 26.9, and 20.2 between observers A and B, A and C, and B and C, respectively, and the RDC was 1.5. Regarding THLD, it was 8.5, 12.7, and 9.4, and the RDC was 1.3. Conclusion Using a standardized methodology, and regardless of the different experiences of the observers, the estimation of THLD is highly reproducible. Although the reproducibility of the assessment of tumor irradiation is overall quite high, large variations may be observed in a limited number of patients.

Intra- and Inter-operator variability in manual tumor segmentation: Impact on radionuclide therapy dosimetry

“…Encouragingly, inter-observer reliabilities are nearly all above 0.8, re ecting substantial agreement between readers, as well. These ndings are consistent with the ndings of Meyers et al [14], who report an inter-observer ICC of 0.94 for volume and 0.73 for mean absorbed dose for delineation on contrast enhanced CT in a similar cohort of HCC patients treated with 90 Y RE. Similarly, McErlean et al [11] determined intra-and inter-reliabilities of 0.957 and 0.954, respectively, for RECIST measurements on CT images, which compare well with our values.…”

“…Such multi-operator studies for manual lesion segmentation are rare in the internal dosimetry setting as it is labor intensive and requires participation by multiple imaging experts. One reported study by Meyers et al [14] investigated inter-but not intra operator variability. Furthermore, that study did not investigate variability associated with individual lesion contours, as the estimated quantity was the dose to the total tumoral liver.…”

Impact of contouring methods on pre-treatment and post-treatment dosimetry for the prediction of tumor control and survival in HCC patients treated with selective internal radiation therapy

“…There have been several approaches to delineate target regions in the literature [ 20 – 24 ]. Most of them have been supporting the idea of delineating targeted regions directly on SPECT/CT or PET/CT images, trying to use the fusion between both images to have a good compromise between the scintigraphic (SPECT or PET) and the morphological volumes (CT) [ 13 , 20 , 21 ].…”

Clinical, dosimetric, and reporting considerations for Y-90 glass microspheres in hepatocellular carcinoma: updated 2022 recommendations from an international multidisciplinary working group

“…Several commercial dosimetry software options are available, streamlining calculation of multicompartment dosimetry. Advancements in catheter technique in addition to refinement of the use of 99m Tc-MAA imaging to estimate Y-90 glass microsphere distribution have led to improved clinical utility [50][51][52][53]. NTAD exhibits better reproducibility, which provides confidence in its use as a safety threshold.…”

“…However, it may be appropriate to screen out patients in whom multicompartment tumor dosimetry predictions may be unreliable; in those cases, defaulting to MIRD is recommended. Recent publications note that variability may be driven by a limited sample size of published data, operator experience, and variability of tumor flow (and hence T:N) in the presence of multifocal disease, suggesting further optimization of patient selection is needed to improve accuracy and reproducibility of multicompartment dosimetry [52].…”