Clinical patient enrollment
During January 2017 to December 2019, hepatic AE patients were evaluated by multidisciplinary team to decide treatment modalities. Inclusion criteria: (i) hepatic AE patients with definitive surgical indications; (ii) patients with both preoperative abdominal CT and 18F-FDG-PET/CT; (iii) patients with enough liver tissue for MSS for pathology during surgery. On the contrary, patients with chronic inflammatory diseases, autoimmune diseases, liver failure (grade Child C), immunocompromised situations, acute cholangitis or hepatitis, severe metabolic disorders were excluded. Subjects with relevant above two imaging examinations but without surgery were included for TBRs analysis.
Patients underwent standardized three-phase CT imaging of the liver, including pre-contrast images plus early and late phase images after contrast medium was introduced venously. Scanning was performed using a 64-slice spiral CT machine (GE Lightspeed Ultra, USA). The voltage and current were 120 kV and 300mA. After routine scanning of the whole body, 40-50 ml of Iohexol (300 mg/ml) was injected intravenously using an Ulrich high pressure syringe at a flow velocity of 4.0-4.5 ml/s and a reservation time of 8 seconds. Continuous scanning was done for 55 seconds with 5 mm slice thickness (layer space was 5 mm). Subsequently, images were transferred to an AW4.2 workstation (GE, USA). All CT images were interpreted as part of the clinical examinations by two experienced radiologists that specialized in abdominal imaging, and a third senior radiological expert was invited for possible bifurcations.
Basic parameters of AE lesions (size and location) and morphological features (density, uniformity, the presence of calcification and its’ characteristics) were recorded in detail. Diameter was measured in three dimensions (axis, coronary and sagittal) and average diameter was calculated as representative. Specially, evaluation of calcification and presence of necrosis was graded with three levels for each: (1) calcification (Cal): (-) no evidence of calcification; (+) minor calcification or central calcification without periphery calcification; (++) diffused or subtotal calcification of the periphery area of the lesion; (2) necrosis (Nec): (-) no evidence of necrosis or uniform lesion; (+) half necrotic half uniform lesion with obvious uniform density area at lesion periphery; (++) almost full necrotic lesion without obvious solid component.
18F-FDG-PET/CT and TBRs
Subjects had whole body 18F-FDG-PET/CT using the Discovery VCT PET/CT (GE Healthcare Bio-Sciences, USA). The tracer 18F-FDG was produced by Cyclotron (GE Healthcare Bio-Sciences, USA) that had a radiochemical purity of >95%. Prior to the examination, patients fasted for at least 6 h to achieve a fasting blood glucose concentration <7 mmol/l. Relevant parameters were: Voltage, 120 kV; tube current, 260 mA; detector collimation, 64×0.625 mm; layer thickness, 3.75 mm; interlayer spacing, 3.75 mm; 0.6 msec/rotation; detector pitch, 0.983; and scanning time, 20-30 sec.
During examination, patients were intravenously injected with 18F-FDG (7.4 MBq/kg body weight) and subjected to take 1,000 ml water orally after 30 min. Further 300 ml water was taken to fill the gut and duodenum following urination with empty bladder. For the next step, positioning images were drawn to determine the scanning range (from the top of the cranium to the mid-upper segment of the thighbone). Patients were educated to breathe smoothly. The three-dimensional PET acquisition was performed with the same scanning range as the CT, generally with 6-8 bed positions. Data for each bed position were collected for 3 min. Then, the CT data were applied to perform attenuation corrections for the PET images. The ordered subset expectation maximization iterative method was used to reconstruct the images of the cross-sectional, coronal and sagittal planes, as well as the PET/CT fusion images.
Imaging diagnosis was carried out by two experienced physicians, and the PET, CT and PET/CT fusion images were independently analyzed: image quality was determined by visual analysis, allowing normal physiological uptake, as well as normal variations and artifacts, to be standardized; the abnormal radiopharmaceutical accumulation and biological borders of the lesions, and the numbers and measurements of the maximum standardized uptake value (SUVmax) accumulated in these lesions were recorded; average SUV (SUVave) of background liver at same slice was measured; according to the anatomical information provided by the same CT machine, the position of each lesion was then precisely identified. TBRs was calculated by SUVmax / SUVave. At last, LME ranges were measured at three different sites to calculate the average scales.
Surgery and medical treatment
Patients were operated for hepatic AE after multidisciplinary team evaluation. Main procedures were conventional hepatectomy (minor, major, excessive), hepatectomy associated with vascular reconstruction, and ex vivo liver resection and autotransplantation for those contraindicated to conventional therapies, and one of them had received auxiliary partial autologous liver transplantation. Post-operatively , all patients were enrolled in regular anti-parasitic medication therapy and followed up routinely according to expert consensus[1, 3].
MSS and pathology
MSS samples were acquired from surgical specimens. The specific sampling site were preoperatively planned through spatial information based on CT and PET/CT, even using three-dimensional reconstruction techniques[13, 14]. First, liver specimens which contained the target lesion (approximately 1 cm at thickness) as well as adherent liver tissue (at least for 5 cm at length) were cut into size about 2cm*3cm*6cm, and fixed with formalin immediately. Note that, not all cases’ liver condition could allow sufficient liver tissue for MSS. For example, (a) left hemi-hepatectomy for left medial lobular lesions could allow MSS due to sufficient left lateral lobular liver parenchyma; (b) left lateral lobectomy for left lateral lobular “full-spaced” lesion could not allow for MSS because of insufficient adherent liver tissue at resection margin. Secondly, after 24-48 h fixation, MSS sampling was performed: liver samples were obtained at parallel levels to lesion surface with 5 mm intervals. Thirdly, paraffin embedded slides were prepared using routine protocols. At last, hematoxylin-eosin (HE) staining and immunohistochemistry for CD3 molecule were performed using manufacturers’ protocols. The widest range was defined by the most distant obviously positive CD3 expression level compared to even more distant slides.
Lesions were categorized into different groups based on calcification (Cal) and necrosis (Nec) metrics (Fig 1):
(A) non-calcified uniform density lesion: Cal (-), Nec (-);
(B) diffuse calcified solid lesion: Cal (++), Nec (-);
(C) half necrotic and half solid with minor calcification: Cal (+), Nec (+);
(D) half necrotic and half solid with obvious calcification: Cal (++), Nec (+);
(E) subtotal necrotic lesion with minor calcification: Cal (+), Nec (++);
(F) total necrotic lesion with obvious calcification: Cal (++), Nec (++);
A demand for at least three repeats (at least three patients) for each lesion types were set ahead of this study to assure the quality of datasets.
TBRs and LME ranges were presented in median with interquartile, and 95% CI were given. One-way ANOVA (Kruskal-Wallis test) and Non-parametric test (Mann Whitney test) were used to determine the significance of TBRs and LME ranges between groups; Non-parametric test (Wilcoxon matched pairs test) was used to compare paired LME ranges indicated by PET/CT and MSS; linear regression was drawn to observe the link of TBRs and LME ranges (r2 was given). P< 0.05 was chosen as standard to judge statistical significance, and actual values were also presented for every test.