Introduction
Multislice computed tomography (CT) integrated positron emission tomography/CT (PET/CT) with Flourine (F)-18 labelled fluorodeoxyglucose (FDG) has been widely used in staging and follow-up of malignant lymphomas as happens in many other solid malignancies in recent years (1,2). Malignant tissues are capable of accumulating more FDG, a radiopharmaceutical of d-glucose analogue, compared to normal tissue due to their excess energy requirement, and represent themselves as highly contrasted foci on PET/CT imaging.
In curable lymphomas, i.e Hodgkin disease (HD) and aggressive non-Hodgkin lymphomas (NHL), the use of standart doxorubicin-containing chemotherapy regimens with or without involved field radiotherapy (IF-RT) has a high success rate. These standart therapy protocols are defined mainly according to histologic subtype and stage of the disease as well as prognostic prediction based on clinical prognostic risk scoring systems (PRS) (3,4). However, a complete long-term disease control can not be achieved in a substantial portion of patients with these standart chemotherapy regimens due to either resistant disease or early relapse. In this subgroup of patients with lymphoma, it is possible to get the disease under control by applying more aggressive chemotherapy protocols (5). However, this approach is required a more serious risk-benefit ratio calculation because of undesired toxic effects of such chemotherapy protocols for the patients (6). Therefore, an effective mid- or end-treatment evaluation is desired in patients with lymphoma and this is of particular importance to discriminate the subgroup of patients either who will have a resistant disease or be at higher risk for early relapse.
In this retrospective study we aimed to determine the efficacy of FDG-PET/CT imaging performed after the first-line treatment in predicting the prognosis of patients with lymphoma and we also compared the PET/CT results with the pretreatment prognostic scores.
Methods
Patients
A total of 103 patients with a diagnosis of histologically proven lymphoma, treated in the hematology clinic of the of the İstanbul University-Cerrahpaşa Hospital, and who underwent FDG-PET/CT scan after the completion of the first-line treatment, were analyzed retrospectively. Our study project was approved by the İstanbul University-Cerrahpaşa Clinical Research Ethics Committee (16.09.2008/27584). There were 56 (54.4%) patients with HD (32 nodular sclerosing, 21 mixed type, and 3 lymphocyte rich) and 47 (45.6%) patients with high-grade NHL (43 diffuse large B-cell lymphoma, 3 anaplastic T-cell lymphoma, and 1 Burkitt’s lymphoma) (Table 1). The age ranged from 16 to 81 with an average of 41.6 ±17.1 (47 females and 56 males). The PET/CT studies were performed between 10 and 90 days after completing their first-line treatment. Thereafter, all patients were followed up in haematology outpatient clinic with an interval of 3-6 months and evaluated by physical examination and laboratory parameters as well as by imaging modalities when needed. Accordingly, their disease outcome status was classified in two groups as “remission” and “non-remission”. The non-remission group was including those patients with partial remission or stable disease, and relapsed or progressed disease. The median follow-up duration was 31.8±7.5 months (ranged from 24 to 55 months). The patients who did not achieve remission were treated by further chemotherapy courses and/or stem cell transplantation with high dose chemotherapy (Table 2a, 2b).
Pretreatment Prognostic Risk Scoring
All patients were staged according to Ann Arbor classification from stage I to IV in the pretreatment evaluation (7). In HD group, stage I-IIA patients were classified as having “early (localised) stage” and stage IIB-IV patients were classified as having “late (advanced) stage. Then, early stage patients were classified in “favorable prognostic” and “unfavorable prognostic” groups according to German Hodgkin Lymphoma study group criteria (8). In NHL group, the patients were classified into low, low-medium, medium-high and high risk groups according to IPI (International Prognostic index) and age-adjucted (aa)-IPI criteria (4). Then, they were categorized within two groups as “low risk” (for low and low-medium groups) and “high risk” (for medium-high and high groups).
First-line Treatment Protocols
According to departmental protocols, patients with early stage HD were treated with ABVD regime from 2 to 4 courses followed by consolidation IF-RT and patients with advanced stage HD were treated with 6-8 courses of ABVD (a total of 10 patients). Nine patients with NHL were essentially treated with CHOP regime with rituximab (only in 9 patients without rituximab), from 3 to 8 courses followed by IF-RT. The indications for consolidation radiotherapy included (i) initial massive mediastinal disease; (ii) individual or confluent nodal masses; and (iii) macroscopic nodules in an intact spleen as determined by CT scan.
PET/CT Imaging and Evaluation Protocol
All patients underwent a post-therapy PET/CT scan, performed 10 to 90 days (median 36.7) after completion of the first-line treatment protocol. A dedicated high-resolution LSO based PET scanner that was integrated with 6 slice CT (Siemens Biograph 6 LSO HI-REZ PET/CT, Illinois, USA) was used for PET/CT imaging. All patients were given an iodine containing contrast agent (Telebrix 30 Meglumin, Guerbet, France or Urovist-angiografin 50, Bayer Türk Kimya, İstanbul, Turkey) diluted in 1.5 L of water and asked to drink it part by part starting 4-8 hrs before their appointment time. Patients were in at least 4 hr fasting at the time of their appointment. After ensuring the peripheral blood glucose level less than 180 mg/dL, 296-703 MBq (8-19 mCi) FDG was intravenously injected into the patient via an IV catheter and thereafter the patients were rested 1-1.5 hr in a semireclining chair in the waiting room. After that, the patients were lain down in supine position on the scanner table. First, a CT topogram image was obtained from the vertex to 1/3 proximal of thighs. Then a low-dose CT imaging (60-80 mAs) without IV contrast administration was performed followed by PET emission imaging in the guidance of the CT topogram. Total PET/CT imaging was generally completed in 7-8 bed positions within 20-25 minutes. Iterative algorithms were applied to both PET and CT data reconstruction using the vendor provided software and multiplanar (axial, sagittal, coronal) slices with approximately 0.5 mm thickness of both PET and CT images as well as maximum intensity projection (MIP) views were obtained. Attenuation correction based on CT data was applied to PET images. Attenuation corrected PET slices together with identically aligned CT slices and coregistered (fused) PET/CT slices as well as MIP images were projected onto high resolution LCD monitors for reviewing and reporting.
The PET/CT images were evaluated by an experienced nuclear medicine phycisian underguidance with all clinical information available. In visual evaluation of PET/CT images, any lymph node with an increased FDG accumulation more than mediastinal blood pool and/or any focus of increased FDG uptake within an extranodal organ was considered as positive finding for residual disease. Additionally, a semi-quantitative index of FDG uptake intensity normalized to the body weight, which was called maximum standart uptake value (SUVmax) was calculated according to standard formula using the vendor-software by drawing a region-of-interest from the most active part of the lesion for each PET-positive region. Additionally, PET-positive patients for residual disease were categorized into 3 groups according to disease sites and extension as follows: i. Single lymphatic station involvement, ii. Multiple lymphatic stations involvement, and iii. Extranodal disease with or without nodal involvement.
Statistical Analysis
The sensitivity, specifity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) of post-treatment FDG-PET/CT imaging and pre-treatment PRS in predicting of disease outcome were calculated for both whole group and subgroups (HD and NHL). Chi-square test was used to compare the diagnostic accuracy rates between the methods, as well as to find out the differences between observed frequencies in the subgroups of the patient population. Kaplan-Meier analysis was used in calculating progression free survival (PFS) curves after first-line treatment for both group and the long-rank test was used for comparison between groups. PFS was defined as the time interval from the end of chemotherapy until progression, relapse, death or date of last follow-up or endpoint of our study. The compatibility between the PET and PRS was evaluated with kappa test. Moreover, the effect of different factors on remission was assessed by Cox regression analysis. SPSS statistical analysis package software (version 15.0) was used for statistical evaluation.
Results
Of the 56 patients with HD, 27 patients had favorable prognostic criteria and remaining 29 patients had unfavorable prognostic status on pre-treatment evaluation. On the other hand, there were 32 patients in low-risk and 15 patients in high-risk categories in the group of NHL patients. Post-treatment PET/CT was negative (no residual abnormality) in 82/103 (80%) patients, and positive in the remaining 21 (20%) patients (Table 3). Of the patients with negative post-treatment PET/CT study, 52 (63%) (22 HD; 30 NHL) had favorable/low-risk PRS, while 30 of them (21 HD; 9 NHL) (37%) had unfavorable/high-risk PRS. In 7 out of 21 (33%) patients (5 HD, 2 NHL) with positive post-treatment PET/CT, there was favorable/low-risk PRS versus unfavorable/high-risk PRS in 14 (67%) patients (8 HD; 6 NHL). There was a significant difference between the favorable/low-risk group and the unfavorable/high risk group in terms of post-treatment PET/CT imaging results (p=0.013). In PET-positive patients, SUVmax values of the lesions ranged from 2.5 to 18.2 (average 9.1).
At the end of follow-up period, 84 of 103 patients (81.6%) stayed in remission. Among these, 52 (61.9%) patients (22 HD; 30 NHL) had favorable/low-risk PRS versus 32 (38.1%) patients (25 HD;7 NHL) with unfavorable/high-risk PRS, while a big majority of the patients (91.6%) had negative PET/CT versus a minority of the patients (8.3%) with positive PET/CT study (Table 4) (Figure 1). Nineteen of 103 (18.4%) patients were in non-remission status, of whom 12 (63.1%) (4 HD; 8 NHL) had unfavorable/high-risk PRS and 7 (36.8%) patients (5 HD; 2 NHL) had favorable/low-risk PRS. In this group, PET/CT was positive in 14 (73.7%) patients versus negative in only 5 (26.3%) patients (Figures 2, 3). Although both PRS and PET/CT imaging had statistically significant capability in terms of estimating remission status of the patients, the statistical power of PET/CT imaging was much stroger than that of PRS (p<0.001 versus p =0.046).
Post-treatment PET/CT imaging demonstrated 74% sensitivity, 92% specificity, 94% NPV, and 67% PPV and 88% accuracy in predicting remission status, whereas for PRS the same values were 63% and 62%, 88%, 27%, and 62%, respectively (p<0.001) (Tablo 5).
Kaplan-Meier survival analysis demonstrated a significant difference for assesment of PFS between PET positive and PET negative patients. (log-rank =2.1 p<0.05), while there was no significant difference between favorable prognostic/low-risk and unfavorable prognostic/high-risk groups for PRS (Figure 4).
There was a compatibility between PET result and remission (kappa =0.62), while no compatibility was available between PRS and remission (kappa =0.17).
On the Cox regression analysis, PET results and disease type were the most significant indices affecting final remission status. When the PET was negative, the remission more likely occurred (p<0.001). The patients with positive PET results have four times more risk to stay in non-remission status. There was also 2.3 fold less remission in the group with NHL comparing to HD group (p<0.05). Those patients with local residual disease on post-treatment PET studies had a more likelihood of staying in remission comparing to those with multifocal lymphatic and/or organ involvement (71% versus 29%; p<0.05) (Table 6).
Discussion
There are numerous reports available in the literature recommending the routine use of FDG-PET imaging to assess the post-therapy response in patients with HD and NHL (9-15). However, there are a considerable variation between the studies in terms of diagnostic accuracy and predictive values most probably due to heterogeneous population studied, difference of treatment protocols, and difference of PET timing etc.
Mid-treatment (interim) FDG-PET/CT scanning performed after 2-4 cycles of chemotherapy appears to have prognostic significance and may have a potential for earlier identification of chemoresistant disease in lymphomas, prior to treatment completion, leading to facilitate an individualized risk-adapted strategy. However, the importance of PET/CT at the end of the first line treatment continues (16-18).
In this retrospective study including a patient population referred from a single clinic, a new generation of hybrid PET/CT scanner was used. We found that post-treatment FDG imaging had 74% sensitivity, 92% specifity, 88% accuracy, 67% PPV and 94% NPV in predicting of long-term remission. These values were significantly better than those of obtained with the pre-treatment PRS, which were 63%, 62%, 62%, 27% and 88%, respectively (Table 5).
Despite PET’s superiority comparing to PRS systems, our study revealed that the FDG-PET/CT was not excellent imaging modality for predicting disease outcome. There were 5 patients with HD and 2 patients with NHL who were were sustained in remission status at the end of follow-up period, despite positive PET results (false-postive rate =33%). This finding was most probably related to inflammatory changes occurring secondary to chemotherapy. Additionally, a sensitive reading of PET images might lead to false-positive results. We used mediastinal blood pool activity as the threshold for the lymph node uptake to define the post-therapeutical PET study as positive. There has been a great effort ongoing to find out what will be the best threshold activity, i.e mediastinum, liver or nearby background to determine positivity on the post therapeutic FDG-PET studies (19). In addition, there are reports available based on a SUVmax threshold or SUVmax difference between the initial study and post-therapeutic studies (20-22). Nevertheless, there has not been defined a well validated interpretation criteria to provide an excellent accuracy for post-treatment FDG-PET studies so far. Nevertheless, there has not been defined a well validated interpretation criteria to provide an excellent accuracy for post-treatment FDG-PET studies so far. Therefore, any ambiguous finding in FDG-PET/CT report that is not supported with clinical-laboratory data or other imaging modalities should be confirmed with biopsy if needed. Particularly, the new lesions that are not available in the pre-treatment PET/CT scan must be evaluated in caution.
On the other hand, negative post-treatment PET/CT scan could not completely predict long term remission, although it had higher NPV. In our group, there were 5 out of 82 (6%) patients (1 HD, 4 NHL) with false-negative PET results, despitesensitive reading. This was most probably due to insufficient spatial resolution of PET imaging technology to detect microscopic residual disease. Therefore, we suggest that the patients in high risk group should be monitored closely even if their post-treatment PET/CT scans are negative.
There was a clear relationship between the disease extention on post-treatment PET scans and relaps rate in our study. The frequency of disease recurrence was significantly higher in the patients with involvement of multiple lymphatic stations or extranodal sites comparing to those with single lymphatic station involvement (Table 5).
Our results are generally compatible with the literature and the values of NPV and PPV are within the limits in published meta-analysis in this field (11,22). There is relatively few studies comparing the prognostic values between post-treatment PET and clinical risk scoring system. In Haioun study, including 90 patients with aggressive NHL, they performed PET studies before treatment, after 2 cycles of chemotherapy and at the end of treatment as well, and prospectively evaluated the predictive value of PET in the prediction of prognosis (23). In this study, 83% of post-treatment PET negative patients, and 58% of post-treatment PET positive patients stayed in complete remission. Two-year PFS was 90% in post-treatment PET negative group and 58% in post-treatment PET positive group. Accordingly, post-treatment PET was an independent prognostic marker in estimating PFS regardless the chemotherapy protocol.
Study Limitations
There are differences in treatment approaches and management of HD and NHL. The main limitation of our study was that it was not performed with a more standard patient group consisting of HD or high-grade NHL. Therefore, we presented both the general data and the data of both groups separately. In addition, the number of patients in both groups was relatively limited and the study was designed retrospectively.
Conclusion
In conclusion, our study confirmed that FDG-PET/CT scanning at the completion of the first-line therapy estimated the disease outcome of the patients with HD and aggressive NHL, better than population based pretreatment clinical risk scoring systems. However, particularly PPV of the end-treatment FDG-PET/CT imaging was still far away from the desired levels as to estimating disease recurrence, due to a relatively high false-positive rate. Nevertheless, our study demonstrated that the patients with persistent FDG uptake in multinodal lymphatic stations and in extranodal sites on the post-therapy PET/CT scanning were under more risk for recurrence. On the other hand, despite higher NPV, there might be some false-negative results with the post-treatment PET/CT imaging for estimation of permanent remission, most probably due to microscopic residual disease available at the end of therapy.
Ethics
Ethics Committee Approval: Our study project was approved by the İstanbul University-Cerrahpaşa Clinical Research Ethics Committee (16.09.2008/27584).
Informed Consent: Retrospective study.
Peer-review: Externally peer reviewed.
Authorship Contributions
Concept: E.B.E., Ş.G., K.S., Design: E.B.E., Ş.G., K.S., Data Collection or Processing: E.B.E., Ş.G., K.S., Analysis or Interpretation: E.B.E., Ş.G., K.S., Literature Search: E.B.E., Ş.G., K.S., Writing: E.B.E., Ş.G., K.S.
Conflict of Interest: No conflict of interest was declared by the authors.
Financial Disclosure: The authors declared that this study received no financial support.