Discussion
PTLD is a proliferative disease that results in abnormal cloning of lymphocytes or plasma cells after solid organ transplantation or stem cell transplantation due to immunosuppression of the recipient, and it can eventually progress to malignant invasive lymphoma. At present, it is believed that the occurrence of PTLD is closely related to postoperative EBV infection(14). The incidence of PTLD after pLT can reach 2%–20%(15-17). In pLT, EBV serological negativity before transplantation is considered to be a major risk factor for PTLD. In adult liver transplantation, there are fewer EBV-seronegative recipients(18), which is one of the reasons why the incidence of PTLD after pLT is higher in children than in adults(19). In addition, another major cause of EBV-related PTLD after pLT is the use of immunosuppressive regimens of high intensity and long duration.
There is limited research on the clinical application of18F-FDG PET/CT in children with PTLD, who are usually combined with adult PTLD cohorts(19-23). The results of an adult cohort showed that 18F-FDG PET/CT is a feasible imaging method for the diagnosis of PTLD. However, the sensitivity values and NPVs of 18F-FDG PET/CT in the detection of PTLD were not high in children(21, 24). At present, there are three problems in the study of 18F-FDG PET/CT in children with PTLD. First, most existing studies were descriptive studies, meta-analyses, or reviews(22, 25, 26) or involved comparisons of other non-pathological examination methods using 18F-FDG PET/CT(27). Second, the numbers of cases in studies on 18F-FDG PET/CT and PTLD in children have been small(28), and there are few studies on PTLD after pLT.
The third problem involves the use of 18F-FDG PET/CT in the diagnosis of PTLD in children. At present,18F-FDG PET/CT has no definite quantitative index in the diagnosis of PTLD in children, and it relies more on the experience of nuclear medicine doctors without a formal guideline(29). When reading the report, the nuclear medicine physician will more or less refer to the patient’s reasons for seeing a doctor and cannot complete the18F-FDG PET/CT diagnosis of PTLD in a completely independent manner. In this study, 18F-FDG PET/CT reports from nuclear medicine doctors for the diagnosis of non-destructive PTLD had high sensitivity values and NPVs, but the specificity values and PPVs were not high. Considering that low uptake of non-destructive or low-grade PTLD may lead to false-negative results, while high uptake related to infection or inflammatory process may lead to false-positive results, the application of SUV in differentiating benign and malignant lesions is still controversial(22, 30). Therefore, finding a quantifiable 18F-FDG PET/CT index to judge PTLD after pLT was the focus of this study.
Unlike traditional tumor-induced lymphadenopathy, our experience found that there was no significant difference in the longest diameter of lymph nodes between PTLD-negative cases and non-destructive PTLD cases. Moreover, the longest diameter of lymph nodes in different parts of the body also varies. Clinically, we are not concerned about the longest diameter of lymph nodes in PTLD patients in the early stage but instead with the morphology of lymph nodes. We found that the lymph nodes of patients with non-destructive PTLD were rounder than those of PTLD-negative cases. This was reflected in the ratio of the shortest diameter of the lymph node to the longest diameter of the lymph node; if the ratio was closer to 1, it meant that the shape of the lymph node was more spherical, while, if it was closer to 0, it meant that the shape of the lymph node was closer to an ellipse. Another problem we found in clinical practice pertains to the SUVmax value. The SUVmax value in our data can effectively distinguish between monomorphic PTLD and polymorphic or non-destructive PTLD. However, although there was a statistical difference in SUVmax between PTLD-negative and non-destructive PTLD cases, the sensitivity, specificity, PPV, NPV, and accuracy obtained using this index were poor. In clinical practice, we have found that there are many reasons that can lead to increased uptake of 18F-FDG in tonsils, posterior pharyngeal wall, or adenoids, such as a high EBV DNA load or tonsil hypertrophy(31, 32). Moreover, 18F-FDG uptake also increases in patients with other conditions after transplantation, such as postoperative inflammation, infection, bone marrow activation, or transplant rejection(19). Therefore, we used the ratio of SUVmaxBio to SUVmaxTon to numerically validate our reasoning in the above clinical practice.
Through our calculation, we found that (SDL/LDL)*(SUVmaxBio/SUVmaxTon) had the largest AUC in each combination, and the cut-off value calculated by this index can effectively judge the non-destructive PTLD in children. Using this cut-off value, the sensitivity, specificity, PPV, and accuracy can reach >90%, and the NPV can also reach 85.7%. However, according to this index, there will still be errors. In the data of our center, there were three false-negative cases and one false-positive case. We found that SUVmaxBio was high in the false-positive case, but the high uptake sites of18F-FDG were concentrated in the neck or supraclavicular lymph nodes, and there was no abnormal uptake of18F-FDG in other parts of the body. In the three false-negative cases, the SUVmaxBio or SDL/LDL was too low, resulting in an incorrect prediction. However, in these three cases, abnormally high uptake of 18F-FDG could be seen in many parts of the patient’s body, including the neck, supraclavicular fossa, armpit, retroperitoneal para-aortic, and inguinal lymph nodes. Therefore, the use of (SDL/LDL)*(SUVmaxBio/SUVmaxTon) and the subjective judgment of nuclear medicine physicians can further improve the accuracy of diagnosing PTLD-negative and non-destructive PTLD cases. It is worth noting that one false-negative case underwent inguinal lymph node biopsy. The inguinal lymph node SUVmax was 1.25, while the cervical lymph node SUVmax was 6.81. If the cervical lymph node SUVmax value was used, the result of this quantitative index would have been PTLD positivity, which was consistent with the pathological results.
It is undeniable that this study has some limitations. First, this study was a retrospective study, necessitating prospective and multicenter studies for internal and external validation in the future. Second, since the gold standard for the diagnosis of PTLD in this study is lymph node biopsy, the selection of lymph node biopsy site became key. The longest and shortest diameters and SUVmax values of lymph nodes involved in this study were those at the biopsy site. Due to clinical limitations, we selected superficial and easily accessible lymph nodes as much as possible. Therefore, some selected lymph nodes were not the most typical lymph nodes used in 18F-FDG PET/CT reports.