Synopsis

Keywords: Cancer, Liver, Hepatoblastoma

This study evaluates the diagnostic efficacy of contrast-enhanced MR imaging for preoperative POSTTEXT staging and prediction of vascular involvement in pediatric hepatoblastoma after neoadjuvant chemotherapy. The findings revealed that MR preoperative POSTTEXT staging has a good level of agreement with the reference standard. MR provides high predictive performance for identifying portal vein involvement and moderate predictive performance for identifying hepatic vein/inferior vena cava involvement.

Purpose

Hepatoblastoma is the most common pediatric liver malignant neoplasm, typically affecting patients aged 6 months to 3 years. Neoadjuvant chemotherapy is commonly used before surgery to improve the rate of complete resection of pediatric hepatoblastoma and reduce tumor recurrence, and the assessment of resectability after chemotherapy is highly dependent on POSTTEXT staging and its annotation factors P (portal vein involvement) and V (hepatic vein or inferior vena cava [IVC] involvement)^1,2. Recently, free-breathing radial stack-of-stars 3D Dixon gradient echo sequence (StarVibe) has enabled the acquisition of high-quality contrast-enhanced MR images in pediatric patients³. It has not yet been documented if MR can be utilized as an accurate preoperative evaluation technique. We aim to assess the diagnostic performance of contrast-enhanced MR imaging for preoperative POSTTEXT staging and prediction vascular involvement in terms of annotation factors P and V in pediatric hepatoblastoma after neoadjuvant chemotherapy.

Methods

This study was approved by the institutional review board with informed consent waived. From September 2021 to September 2022, a total of 32 consecutive pathologically proven pediatric hepatoblastoma patients who had MRI following neoadjuvant chemotherapy with a time interval between MRI and surgery of fewer than 14 days were collected retrospectively (Fig. 1). MRI was performed on a 3T scanner (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany), with coronal T2w Half-Fourier acquisition single-shot turbo spin-echo (HASTE), axial T2w HASTE fat-suppression, axial DWI. For dynamic imaging, axial T1w StarVibe Dixon (Trajectory = Radial, FOV = 300 mm × 300 mm, TR/TE = 2.65/1.42 ms, NSA = 1, Flip angle = 9 deg, GRAPPA = None, Bandwidth = 980 Hz/px, Radial views = 384, voxel size = 1.2 mm × 1.2 mm × 2.5 mm) was performed before and 0 s, 49 s, and 137 s after contrast administration of 0.1 mmol of gadopentetate gadobutrol (Gadavist, Bayer Healthcare) per kilogram of body weight at a rate of 1 mL/s with saline flush. Coronal Cartesian T1w Vibe Dixon (Trajectory = Cartesian, FOV = 280 mm × 350 mm, TR/TE/TI = 4.22/1.32/2.55 ms, NSA = 6, Flip angle = 9 deg, GRAPPA = CAIPIRINHA:2, Bandwidth = 980 and 750 Hz/px, voxel size = 0.6 mm × 0.6 mm × 2.5 mm, Radial views = None) was performed 98 s after contrast. The entire protocol takes about 20 minutes from the localizers to the end of the scan. Two radiologists jointly reviewed the MR images and recorded POSTTEXT stages as well as annotation factors P, V, E (extrahepatic disease contiguous with the main liver tumor), C (caudate involvement), F (multifocality), N (lymph node metastases), R (tumor rupture), M (distant metastases) based on 2017 PRE/POSTTEXT system. The combination of operational records and pathologic findings served as the reference standard. Preoperative MR POSTTEXT staging was compared with the reference standard using the Kappa test. Diagnostic performances of MR in assessing P and V status were evaluated using the receiver operating characteristic (ROC) curve. The discrepancies between MR POSTTEXT and reference standard were further analyzed.

Results

Among 32 patients, 5 (15.6%) were POSTTEXT stage I, 15 (46.9%) were stage II, 6 (18.8%) were stage III, 6 (18.8%) were stage IV. Preoperative MR POSTTEXT staging was correct in 20 patients (62.5%, weighted Kappa 0.61; 95% CI, 0.42 to 0.80). Four patients had portal vein involvement and four had hepatic vein/IVC involvement (Fig. 2 and 3). The sensitivity, specificity, and AUC were 100%, 96.4%, and 0.982 for portal vein involvement and 75.0%, 75.0%, and 0.750 for venous/IVC involvement, respectively. Overstaging was present in four (12.5%) patients (two cases overstaged I to II, and two cases II to III), and understaged in eight (25.0%) patients (two cases understaged II to I, three III to II, and three IV to III). The reasons for these 12 mis-staging cases including difficulty in determining the relationship between lesion boundaries and the landmark vessels (n = 4) (Fig. 4), discrepancy of caudate lobe boundary between radiologist and surgeon (n = 3) (Fig. 5)⁴, microlesions (n = 2), pedunculated tumor (n = 1), large tumor bulging landmark vessels (n = 1), and duplicated left hepatic vein (n = 1).

Conclusion

A 20-minute free-breathing MRI contrast-enhanced examination for preoperative POSTTEXT assessment is feasible in pediatric hepatoblastoma patients. Preoperative MR POSTTEXT staging has a good level of agreement with surgical and pathological findings. MR has excellent accuracy for assessing P-status and moderate accuracy for V-status. The non-excellence performance of MR in assessing V status is most likely due to StarVibe's inadequate resolution in the z-axis (2.5mm), through which the hepatic veins and inferior vena cava travel. As StarVibe does not yet support coronal scans, coronal contrast-enhanced images using Cartesian trajectory remain susceptible to motion artifacts.

Acknowledgements

No acknowledgement found.