Comparison of the quantitative measurement of 18F-FDG PET/CT and histopathological findings in IgG4-related disease

Although there is increasing use of 18F-FDG PET/CT in the diagnostic procedure for IgG4-related disease (IgG4-RD), little is known about the quantification of 18F-FDG PET/CT. We aimed to evaluate the utility of the quantification of 18F-FDG PET/CT in the diagnosis of IgG4-RD by analyzing the relation between 18F-FDG PET/CT findings and histopathologic findings. the biopsy site in IgG4-RD. The calculation of SUV mean , not of SUV max , is important for evaluating IgG4-RD-related lesions in 18F-FDG PET/CT imaging.


Introduction
IgG4-related disease (IgG4-RD) is a systemic fibroinflammatory condition characterized by IgG4-positive plasma cell infiltration of the affected organs or tissue resulting in organ enlargement and organ fibrosis 1,2 . The commonly affected organs are the lacrimal glands, salivary glands, periorbital tissue, pancreas, lymph nodes, retroperitoneal tissue, lung, kidney, and skin, and dysfunctions of these organs are sometimes caused as a consequence 3,4 . The diagnosis of IgG4-RD is based on the combination of characteristic histopathologic, clinical, serologic, and radiologic findings 5,6 . IgG4-RD should also be differentiated from multiorgan disorders that may present organ enlargement, lymph node swelling, and tumor-like swelling, such as lymphoproliferative disease and neoplastic disease 6,7 . Thus, it is important to biopsy the organ or tissue that is suspected to be affected by IgG4-RD not only for diagnostic but also differential diagnostic purposes. However, until now, little has been known about the method of selecting the biopsy site from multiple suspected lesions. Also, depending on the organ involved, e.g. the aorta or pituitary gland, a biopsy is sometimes a difficult and invasive procedure. Thus, a non-invasive method that can provide useful information for selecting a biopsy site that is suspected of being IgG4-RD-associated from among the lesions is needed.
18 F-fluoro-deoxy-glucose positron emission tomography/computed tomography ( 18 F-FDG PET/CT) is a functional imaging procedure that is widely used in the diagnosis, staging, treatment response, and relapse monitoring of various types of malignancies, including malignant lymphoma 8,9 . It may be considered a valuable tool in the workup of patients with newly diagnosed IgG4-RD, in particular for the detection of multiple organ involvement 10-12 . Additionally, there is growing evidence that 18 F − FDG uptake is correlated with the treatment response 10,13,14 .
Quantitative analysis of PET parameters is a more accurate, objective, and reproducible means of assessing PET images. The utility of the quantification of 18 F-FDG PET/CT for the differential diagnosis of patients clinically suspected of having IgG4-RD had been reported 15-17 . In those studies, 18 F-FDG uptake was lower in IgG4-RD than in malignancies 15,16 . However, the appropriate method of quantifying 18F-FDG PET/CT may differ by disease, and a suitable method for use in Ig4-RD patients remains unknown. To identify or suspect the IgG4-RDassociated lesions among multiple lesions with 18 F-FDG uptake using a semiquantitative method, an appropriate method for quantification is needed. Such a method would contribute to the selection of a biopsy site for diagnosis. Here, we evaluated the utility of quantifying 18 F-FDG PET/CT in the diagnosis of IgG4-RD by analyzing the relation between 18 F-FDG PET/CT findings and histopathologic findings. In addition, we determined the optimal method for the quantification of 18 F-FDG-PET/CT in IgG4-RD.

Patients
Patients were recruited from the Department of Immunology and Rheumatology, Nagasaki University Graduate School of Biomedical Sciences, Isahaya General Hospital, and Nagasaki Medical Center. A total of 21 consecutive patients, in whom 18 F-FDG PET/CT was undertaken at the time of diagnosis between November 2011 and July 2018, were enrolled. All patients fulfilled the comprehensive diagnostic criteria for IgG4-RD 5 and 2019 American College of Rheumatology(ACR) /European League Against Rheumatism(EULAR) classification criteria for IgG4-RD 6 . The decision as to whether to undergo 18 F-FDG PET/CT in order to exclude other diseases or to evaluate the organ involvement of IgG4-RD were made by each physician with the agreement of each patient. Tissue biopsies for at least one lesion with abnormal 18 F-FDG uptake were performed in all patients during the diagnostic procedure. The patients gave their informed consent to be subjected to the protocol, which was approved by the Institutional Review Board (IRB) of Nagasaki University (IRB approval no.: 17091109).
We retrospectively reviewed the clinical, laboratory, and histopathological features and findings of 18 F-FDG PET/CT and evaluated the clinical course for 1 year.

Histopathological findings
Tissue biopsy was performed at the time of diagnosis. According to pathologic criteria for diagnosis of IgG4-RD proposed by the comprehensive diagnostic criteria for IgG4-RD and 2019 ACR/EULAR classification criteria for IgG4-RD 5,6 , we defined tissue that exhibited the following histopathological characteristics as histopathologypositive tissue. These histopathological characteristics include marked lymphocyte and plasmacytic infiltration with fibrosis, infiltration of IgG4-positive plasma cells with the ratio of IgG4+/IgG + cell > 40%, and an elevated number of IgG4 + plasma cells per high power field (> 10).

F-FDG PET/CT imaging
All patients underwent a 18 F-FDG PET/CT scan on a dedicated PET/CT system (Siemens Biograph mCT, Germany, or Siemens Biograph 16true point, Germany, or GE Discovery ST, Wisconsin, USA) consisting of a PET scanner and a 64-row multidetector CT scanner. The patients fasted for more than 5 hours before the intravenous injection of 18 F-FDG. 18 F-FDG was injected intravenously at 180-325 MBq. One hour after the 18 F-FDG injection, the scanning was performed from the middle of the thigh to the top of the skull. The CT data were used for calibration (attenuation correction). 18 F-FDG uptake was assessed at the site of major organ involvement of IgG4-RD, which could be differentiated from the normal uptake of background tissue with 18 F-FDG PET/CT.

Image analysis
For quantitative analysis, PET/CT images were extracted. We measured the highest standardized uptake value (SUV) of the pixels (SUV max ) and the average SUV (SUV mean ) within the biopsied tissue using Metavol, which is a dedicated open-source software ( Fig. 1) 18 . We also measured the SUV mean of the liver as a reference. Then, we calculated the ratio between the SUV mean of the biopsied tissue and the SUV mean of the liver (SUV mean /liver).

Statistical analysis
GraphPad prism software (GraphPad Software, San Diego, CA) and JMP Statistical Software (SAS Institute, Cary, NC) were used for the statistical analysis. We used the Kolmogorov-Smirnov test to check the normal distribution, and an unpaired student's T test was used to detect statistically significant differences in the value of SUV. Receiver operating characteristic (ROC) curves were constructed to establish the cut-off values of SUV mean and SUV mean /liver on which to base biopsy decisions. P-values less than 0.05 were considered to indicate statistical significance.

Baseline characteristics of the patients and disease course
The subjects' characteristics are listed in Table 1. There were 21 patients (11 males, 10 females).

Quantitative 18 F-FDG PET/CT findings and histopathological findings
Tissue biopsy was performed at 24 sites (lymph node 4, submandibular gland 10, prostate gland 4, pancreas 2, thyroid gland 1, lung 1, retroperitoneum 1, and kidney 1) in 21 patients. Among these 24 sites, 19 tissues were histopathology-positive, whereas the histopathological findings for the other 5 tissues were not consistent with the pathologic criteria for the diagnosis of IgG4-RD (Table 2). These 5 histopathology-negative tissues were also not consistent with other diseases such as malignant lymphoma. Almost all of these histopathological findings showed lymphocyte and plasmacytic infiltration with few/no IgG4-positive cells. The values of SUV max and SUV mean for each organ and SUV mean /liver are summarized in Table 3. The IgG4-RDinvolved organ with the highest SUV max was the lung, and the lowest was the retroperitoneum (7.61 and 3.37, respectively), and the same tendency was seen in the analysis of SUV mean (6.04 and 3.06, respectively).
However, in the analysis of the ratio between SUV mean and SUV mean of the liver, the organ that showed the highest value was the lymph node, while the prostate showed the lowest (2.48 and 1.43, respectively). There were no significant differences in the values of SUV max , SUV mean and SUV mean /liver among the organs. The mean values of SUV max , SUV mean , SUV mean /liver in all sites were 5.26, 4.68 and 2.04, respectively.
Table3. Biopsy site and the value of standardized uptake value (SUV). Data are mean ± standard deviation (SD). SUV mean /liver was the ratio between the SUV mean of the biopsied tissue and the SUV mean of the liver.

Comparison between 18 F-FDG uptake and histopathological findings
We next examined whether 18 F-FDG uptake was correlated with histopathological findings. Although there was no significant difference in the value of SUV max between histopathology-positive and histopathology-negative tissues, the values of SUV mean were significantly higher in the histopathology-positive tissues (Fig. 2). The values of SUV mean /liver were also higher in those of the histopathology-positive group as compared to in the histopathology-negative tissue.
To calculate the cut-off value of SUV mean to determine which among multiple lesions is to be biopsied, a ROC curve was constructed (Fig. 3A). Curves were drawn for SUV mean and SUV mean /liver, and the areas under the curves (AUCs) were 0.81 and 0.82, respectively. According to ROC curve analysis, at the best discriminative SUV mean cut-off of 4.07, the sensitivity and specificity were 79.0% and 80.0%, respectively. Moreover, ROC curve analysis found SUV mean /liver = 1.66 to be the cut-off value with a sensitivity and specificity of 84.2% and 80.0%, respectively (Fig. 3B)

Discussion
We investigated the relation between the value of SUV on 18 F-FDG PET/CT imaging and the histopathological findings of suspected IgG4-RD-involved lesions. To avoid unnecessary biopsies and to select suitable lesions for biopsy in diseases with multi-organ involvement such as IgG4-RD, determining which lesion is suspected to be disease-involved using a non-invasive test is important. The results of this study indicated that the value of SUV mean , not SUV max , has good diagnostic performance in IgG4-RD.
In IgG4-RD, 18 F-FDG PET/CT has been reported to be more accurate and more sensitive for detecting the lesions To analyze the characteristics of suspected IgG4-RD-involved lesions, quantitative analysis was performed. Additionally, to identify the optimal quantitative method for IgG4-RD, we also performed three methods of quantitative analysis. The optimal method for the quantification of PET/CT imaging might be different depending on the disease. For the differentiation of metastatic lesions of non-small-cell lung carcinoma from benign lymph nodes, the SUV max -lymph node/ SUV max -primary tumor ratio showed better diagnostic performance as compared to SUV max -lymph node 19,20 . Another study showed that the median SUV was correlated with local recurrence of a squamous cell carcinoma of the anal canal but that SUV mean was not correlated with local recurrence 21 . In patients with head-and-neck cancer, SUV mean , not SUV max , was correlated with disease-free survival, and so SUV mean seemed to have the potential to become a prognostic factor in head-and-neck cancer 22 . Our present study showed that SUV mean and SUV mean /liver were significantly correlated with biopsyproven IgG4-RD involvement, but that SUV max was not correlated. This is consistent with the results of other studies. Zhan et al. reported that diffusely elevated 18 F-FDG uptake in organs should be considered to be more likely to indicate IgG4-RD lesions as compared to a patchy/nodule uptake pattern 12 . Diffused and enlarged tissues/organs are clinicopathologic characteristics of IgG4-RD 23,24 , such as in the following examples: 1) diffuse infiltration of sinonasal mucosa with an IgG4-positive plasmacytic infiltrate in IgG4-reralted chronic rhinosinusitis; 2) diffusely thickened gastric mucosa on endoscopy in IgG4-related gastritis; 3) diffusely enlarged sausage-shaped pancreas in autoimmune pancreatitis (considered to be on the IgG4-RD spectrum); 4) homogenous thyroid grand enlargement in IgG4-related thyroiditis; and 5) diffuse kidney enlargement and diffuse immune complex deposition in the tubular basement membrane in IgG4-relatad kidney disease. These characteristics are also consistent with our study; namely, the comparison of SUV mean between tissues with abundant IgG4-positive plasmacytes and tissues with few/no IgG4-positive plasmacytes in our study showed more significant differences as compared to the analysis of SUV max . SUV max is the most commonly used 18 F-FDG PET/CT parameter in daily clinical practice, and it is often included in imaging reports. It is simple and quick to measure, but SUV max may not represent the status of ROI in IgG4-RD because its value is taken from a single voxel. SUV mean within an ROI is supposed to represent disease status in IgG4-RD more precisely, as we found in this study; therefore, we should calculate SUV mean to identify IgG4-RD lesions. Furthermore, according to ROC curve analysis, the AUC using the value of SUV mean /liver was higher as compared to that using the value of SUV mean . This suggested that SUV mean was influenced by the individual hepatic metabolism, and SUVmean/liver might be more preferable.
ROC curve analysis also indicated SUV mean =4.07 or SUV mean /liver = 1.66 as the cut-off value that discriminated IgG4-RD-associated lesions. This value might be useful for selecting appropriate tissue for biopsy among multiple lesions that are suspected to be associated with IgG4-RD by 18 F-FDG uptake or to determine if an 18 F-FDG uptake lesion on which a biopsy cannot be performed is an IgG4-RD-associated lesion. However, this cut-off value should be interpreted with caution because the appropriate cut-off value might be different in each organ. For example, the SUVs in the prostate and pancreas were relatively low in our study, although the SUV values of those organs have been reported to be almost the same in normal subjects as compared to other organs 25,26 .
The appropriate cut-off value for each organ should be identified by analyzing a larger number of patients in the future.
Several limitations of this study must be mentioned. The number of patients was small, and the longitudinal analysis was not fully performed. We could evaluate the clinical course for one year in 11 out of 21 patients, but there were no relapses or progression of disease. We therefore could not compare the value of 18 F-FDG PET/CT at baseline with respect to the treatment response or disease progression. To make PET/CT imaging more useful for the management of treatment for IgG4-RD, analysis of the relation of the SUV value with the treatment course such as an inadequate response to steroid treatment, the rate of relapse and concomitant use of immunosuppressant desistance is needed in a larger longitudinal study. The values of SUV mean of lymph node were relatively higher as compared to other tissues, and all histopathologic findings of lymph nodes were histopathology-positive in our present study. These results influenced the cut-off value in this study. The value of SUV mean of histopathology-negative lymph node should be compared with that of histopathology-positive lymph node, and mentioned above in the discussion section, further analysis is needed of the cut-off value of SUV mean focused on each organ using larger samples. Also, comparison of the SUV mean of IgG4-RD with those of other diseases such as malignant lymphoma was not done in this study. The utility of the quantification of PET/CT for differentiating IgG4-RD from other diseases should be assessed in a future study.

Conclusions
In conclusion, our present study suggested that quantitative analysis of 18 F-FDG PET/CT imaging is useful for selecting the biopsy site in IgG4-related disease. The calculation of SUV mean, not of SUV max , is important for evaluating IgG4-RD lesions in 18 F-FDG PET/CT imaging.