F-18 FDG PET-CT Versus Contrast Enhanced CT in Detection of Extra Nodal Involvement in Patients with Lymphoma.

AIM AND
OBJECTIVES: The aim of this study is to assess the added value of hybrid fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) in the evaluation of extranodal involvement in patients with lymphoma in comparison to contrast-enhanced CT (CECT). PATIENTS AND METHODS: All patients had lymphoma, proved by histopathological and immunophenotyping examinations. They underwent CECT and F-18 FDG PET-CT studies. Both CECT and PET/CT studies were done within 30 days.
RESULTS: The study included 144 patients 92 (63.9%) males and 52 (36.1%) females with mean age 49.3 years (range 18-80 years). A total of 102 (70.8%) patients had non-Hodgkin lymphoma and 42 patients (18.1%) had Hodgkin disease, diffuse large B-cell lymphoma subtype had the highest prevalence 52.8% (76/144), whereas lymphocytic predominance was the least prevalent 1.4% (2/144) followed by mucosa-associated lymphoid tissue lymphoma 2.8% (4/144) and small lymphocytic type 4.2% (6/144), mixed cellularity (MC), T-cell, and follicular type were equally distributed 6.9% (10/144 each). The lung was the most common site as it was involved in 34 patients followed by bone and bone marrow 32 patients, spleen 18, liver 16, nasopharynx 8, stomach 6, cutaneous and subcutaneous tissue 6, peritoneum, cecum, small intestine, brain, and intramuscular four patients each. However, the parotid and pancreas were the least common sites two patients each. The overall sensitivity, specificity PPV, NPV, and accuracy of PET/CT and CECT are 97%, 20% 94.2%, 33.3% and 91.7% and 89.6%, 60%, 96.8%, 30% and 87.5%, respectively.
CONCLUSION: F-18 FDG-PET/CT can accurately monitor the extranodal lymphoma, it can detect metabolically-active lesions without CT structural changes and identify viable tumor in normal size lymph nodes. FDG-PET/CT is more effective than CECT in evaluating extra nodal lymphomatous infiltration, especially in the spleen, bone, and bone marrow.

Introduction

Lymphoma comprises a histologically heterogeneous group of cancers derived from the cells of the immune system.[1] It accounts for approximately 5%–6% of all malignancies.[2] Lymphoma, mainly non-Hodgkin's lymphoma (NHL), maybe extranodal in 40% of patients,[3] the term extranodal lymphoma has been used to describe this form of lymphoid malignancy, in which there is neoplastic proliferation at sites other than the expected native lymph nodes or lymphoid tissue.[3] The observed rising incidence of NHL and Hodgkin disease (HD) in the past two decades has been characterized by a marked increase in the occurrence of extranodal lymphoma.[4] Lymphomas that initially appear to have the bulk of the disease at extranodal sites are described in primary extranodal lymphoma and categorized as Stage I or II. In secondary extranodal lymphoma, there is secondary involvement of the extranodal sites from primary nodal disease, which is categorized as Stage III or IV.[5] Cross-sectional anatomical imaging techniques, particularly CT, have been the primary modality for the diagnosis, staging, restaging, and follow-up of patients with lymphoma. However, these modalities have several limitations when detecting nodal or extranodal disease, because CT is based only on anatomical structural changes.[5] Fluorodeoxyglucose (FDG) positron emission tomography (PET) imaging has been shown to be an important technique for both staging and follow-up of nodal and extra nodallymphoma.[6] PET–computed tomography (PET–CT) systems, which enable the performance of PET and CT data acquisition at the same setting without changing the patient's positioning, have been recently introduced in clinical practice. Lesions are characterized on the fused PET–CT images by both their metabolic status and their anatomic details. Such fusion can also assist in the differentiation of physiologic and pathological sites of FDG uptake.[7]

Aim of the work

The goal of this study is to assess the diagnostic impact of hybrid FDG-PET/CT in the detection of extranodal involvement in various structures and organs in patients with lymphoma compared to contrast-enhanced CT (CECT) in correlation with pathological data.

Patients and Methods

The protocol of this study was approved by the Ethical Committee of the Board of Nuclear Medicine and Radiology at the National Cancer Institute (NCI) and the International Medical Center (IMC). The study was conducted during the period from June 2014 to September 2017. The study included 144 patients (92 males and 52 females). All patients had lymphoma, diagnosed initially by pelvi-abdominal CECT and proved by histopathological and immunophenotyping examinations. The patients were collected from the NCI, IMC and Zagazig University hospitals. All patients gave informed consent for study participation before imaging. The patients underwent CECT, whole-body FDG PET–CT, and histopathological examination of sites of extranodal involvement. PET/CT studies were carried out at the PET/CT units of the NCI and IMC. FDG-PET/CT and CECT studies were performed within 1 month.

Contrast-enhanced computed tomography scan

Head, neck, chest, and abdomen CT scan was done on 64 multi-detector CT scanner using nonionic iodinated contrast in a dose of (2.0 ml per kilogram body weight) that was injected via ante-cubical vein with an overall injection time of 32 s by the automated injector. All contrast-enhanced CT studies were interpreted by radiologists.

Positron emission tomography/computed tomography scanning and image analysis

FDG PET/CT scan was done using PET/CT scanner (Philips PET/CT Gemini TF 64 multislice CT detector). This device merge a PET scanner with a multidetector CT scanner and permit the acquirement of co-registered PET and CT images in a single session. All patients fasted for 4–6 h before the administration of 370 MBq 18 F-FDG. Imaging started 60 min following tracer injection (5–7-bed positions; acquisition time, 2–3 min/bed position). Blood glucose level did not exceed 150 mg/dL. Initially, patients were imaged in the supine position with arms behind head; CT scanning begins at the level of skull base using the following parameters: 40 mAs; 130 kV; slice thickness, 2.5 mm; pitch, 1.5. The CT scans were done during shallow normal breathing and reached caudally to the mid thighs. PET was performed over the same region at once after gaining of the CT images. Attenuation correction was done using CT-data and images were reconstructed as 3-mm slices applying a standard iterative algorithm (ordered-subset expectation maximization). Images were interpreted at a workstation equipped with fusion software that provides multi-planar reformatted images and enables display of the PET images, CT images, and fused PET/CT images in any percentage relation. Side-by-side image interpretation was accomplished by two experienced nuclear medicine physicians. The analysis was performed using a multimodality computer platform. For semi-quantitative analysis, the nuclear medicine physician referred to the PET/CT fusion images and the CT images to set a spherical volume of interest (VOI) over the regions of interest and then recorded the peak standardized uptake value (SUVmax) in the VOI. PET/CT images were analyzed by a skilled radiologist and a qualified nuclear medicine physician.

Statistical analysis

Data collected through history, basic clinical examination, laboratory investigations, and outcome measures coded, entered, and analyzed using Microsoft Excel software. Data were then imported into Statistical Package for the Social Sciences (SPSS 20, IBM, Armonk, NY, United States of America) software for analysis. According to the type of data qualitative data represented as number and percentage, quantitative continues group represented by mean ± standard deviation, the following tests were used to test differences for significance difference and association of qualitative variable using Chi-square test. Differences between parametric quantitative independent groups by t-test in nonparametric by Mann–Whitney, multiple analyses by Kruskal–Wallis. The correlation was also done using Pearson's or Spearman's, Kappa agreement for agreement. P value was set at <0.05 for significant results and <0.001 for high significant result.

Results

The study included 144 patients 92 (63.9%) males and 52 (36.1%) females with mean age ±14.56 years (range 18–80 years) all patients had histologically proven lymphoma 70.8% (102/144 patients) NHL and 29.2% (42/144 patients) HD. Extranodal involvement was histopathologically proven in 93% (134/144 patients), diffuse large B-cell lymphoma (DLBCL) subtype had the highest prevalence 52.8% (76/144) while lymphocytic predominance was the least prevalent 1.4% (2/144) followed by mucosa-associated lymphoid tissue (MALT) and small lymphocytic lymphoma 2.8% (4/144) and 4.2% (6/144), respectively, while, MC, T-cell, and follicular lymphomas were equally distributed 6.9% (10/144 each) [Table 1]. The most common extranodal sites were the lung, bone, and bone marrow (34and 32 patients each) followed by spleen 18, liver 16, stomach, cutaneous and subcutaneous tissue 6 patients for each sit while nasopharynx, peritoneum, cecum, brain, and intramuscular had 4 patients for each. The parotid, small intestine and pancreas were the least common sites 2 patients each [Table 2 and Figure 1]. Although, extranodal lymphoma was more common in the age-groups 50–60 and 60–70 years than the other age-groups it was not statistically significant (P = 0.455).

Table 1

Distribution of pathological types and subtypes

Table 2

Relation between standardized uptake value and sites of extra nodal involvement

Figure 1

Distribution of extranodal sites

PET/CT was positive in 138 (95.8%) patients out of them 130 were true positive, while CECT was positive in 124 (86.1%) patients out of them 120 were true positive. PET/CT was negative in 6 (4.2%) patients with only two true negative compared to 6 true negative in CECT that showed 20 (13.9%) negative patients [Figure 2, Tables 3 and 4]. PET/CT had higher overall sensitivity than CECT 97% (130/134) versus 89.6% (120/134), but this difference is not statistically significant (P = 0.55), whereas the later had higher specificity 60% than the former 20% with statistically significant difference (P = 0.000). The PPV, NPV, and accuracy of PET/CT are 94.2%, 33.3%, and 91.7% compared to 96.8%, 30%, and 87.5% for CECT [Tables 3 and 4]. SUVmax varied among the different histopathological subtypes, the T-cell lymphoma had the highest mean values (13.2 ± 5.76), and the lymphocytic predominance had the lowest values (6.0 ± 00).

Figure 2

Overall findings of positron emission tomography/computed tomography, computed tomography + C and pathology

Table 3

Positive predictive value, negative predictive value, sensitivity and specificity and accuracy of positron emission tomography/computed tomography and contrast enhanced computed tomography

Table 4

Analysis of the positron emission tomography/computed tomography and CT + C results in correlation with the pathological finings

DLBCL type showed the highest range value (range 1.3–36 and mean 9.49 ± 7.52) with moderate uptake in most of the patients, nodular sclerosis (NS) and cutaneous T-cell types showed also moderate uptake while, small lymphocytic and lymphocytic predominance showed low uptake, however, MALT and follicular lymphomas showed low-to-moderate uptake (range 2.6–23.0 and 1.00–11.4, respectively). SUVmax could not be used to differentiate between the different subtypes as no statistically significant difference could be seen P = 0.38 [Table 5].

Table 5

Relation between standardized uptake value and pathological type and subtypes

The highest mean SUV of FDG uptake was in lesions seen in the muscles, caecum and the spleen (mean SUVmax. 17.55, 16.65, and 12.9, respectively) while the lowest uptake was in lesions seen in the pancreas, the small intestine and the parotid gland (mean SUVmax. 2.70 ± 0.00, 4.65 ± 0.75 and 4.65 ± 0.75, respectively) while the other sites showed moderate uptake values [Table 5, Figures 3 and 4].

Figure 3

Metabolically, active gastric lymphoma together with lymph nodes involvement. Axial multi-slice contrast-enhanced full dose computed tomographyof the upper abdomen (a), axial positron emission tomography (b) and axial positron emission tomography/computed tomography (c) images show an irregular circumferential pyloric mural thickening seen with a metabolically active soft tissue mass lesion measuring 7.2 cm × 4.6 cm with standardized uptake value maximum 23.8, the stomach appeared hypotonic as well, few glucose avid left gastric lymph nodes seen with the largest one measures 13.5 mm × 13.5 mm with standardized uptake value max 3.6

Figure 4

Coecal and brain extranodal lymphoma. Axial: Contrast enhanced computed tomography (a), positron emission tomography (b) and positron emission tomography/computed tomography (c) an ill-defined enhancing lesion in the right frontotemporal region with increased tracer uptake (lesion/nonlesion ratio) = 2.3 (n < 1.5). Coronal: Contrast-enhanced computed tomography (d), positron emission tomography (e) and positron emission tomography/computed tomography (f) coecal circumferential metabolically active wall thickening with smudging of the fat planes (standardized uptake value maximum 13.3) together with a metabolically active right external iliac lymph node with standardized uptake value max 8. Sagittal: Contrast-enhanced computed tomography (g) and positron emission tomography (h) diffuse increased tracer utilization in the whole axial skeleton and spleen denoting bone marrow hyperplasia

Discussion

F-18 FDG PET imaging (18FDG-PET/CT) has been shown to be an important technique for both staging and follow-up of nodal and extranodal lymphoma.[6] PET–CT systems, which enable the performance of PET and CT data acquisition at the same setting without changing the patient's positioning, have been recently introduced in clinical practice. Lesions are characterized on the fused PET–CT images by both their metabolic status and their anatomic details. Such fusion can also assist in the differentiation of physiologic and tumoral sites of FDG uptake.[7] The study showed that lymphomas are more common among males 63.9% (92/144) than females 36.1% (52/144), this comes in agreement with Ömür et al. whose study showed males are more affected than females.[5] In contradict, El-Haddad et al. study showed that the disease is less common in males than females.[8] In the current study, DLBCL was the most common pathological subtype 52.8% (76/144) followed by NS 18.1% (26/144) while lymphocytic predominance and MALT were the least common subtypes 1.4% and 2.8%, respectively, in agreement with Omar et al. study in which DLBCL was the most common 43.6% (48 patients) followed by NS 20% (22 patients),[9] also these results come in agreement with those of Raanani et al.who showed that DLBCL represent 33% of the whole patients included in the study and NS represent 28.2% of the patients.[10] Furthermore, our results are nearly in line with Das et al.who concluded that DLBCL is the commonest NHL type representing 44.1% of total extranodal lymphoma followed by follicular lymphoma and NS is the most common HD type followed by MC.[11] In the present study, the avidity of different histological subtypes, as well as the same subtype, varied widely which is consistent with Ngeow et al. and Parra et al. who showed a very wide range of FDG uptake from low-to-very high (SUVmax1.3–36) seen in DLBCL and NS subtypes.[1213] MALT lymphoma had moderate avidity while small lymphocytic lymphoma showed low-to-moderate avidity that is in line with Parra et al. who showed none to low uptake in the later subtype that was explained by Parra et al.as FDG avidity is multifactorial process. Hutchings et al.[14] investigated SUV levels in the different histopathological subtypes of HL and found that the mean SUVmax varies among the different subtypes. The Mean SUV was 9.3 g/ml in nodular lymphocyte predominance (NLP) patients, 16.3 g/ml in NS patients, 20.8 g/ml in MC patients, which disagreed with our study that showed mean SUV 6.0 g/ml in NLP patients, 8.4 g/ml in NS patients and 7.0 g/ml in MC patients. This study found that the most common sites for extra nodallymphomatous infiltration were the lung followed by bone and bone marrow, spleen, liver, gastrointestinal tract (GIT) (stomach, small intestine, and cecum), head and neck (nasopharynx and parotid) in decreasing order, while the pancreas and the parotid glands were the least commonly affected sites. This is to some extent coinciding with the results reported by Ömür et al.[5] who stated that the most common extranodal sites are as follows: lung, bone and bone marrow, Waldeyer ring, spleen, and liver.

The parotid is the only salivary gland that was involved that comes in agreement with Metser et al.[3] who stated that parotid gland is the most commonly involved salivary gland. In decreasing order, the abdominal organs involved in this study were as follows: spleen, liver, GIT, peritoneum, and pancreas that is nearly in line with Lee et al.[15] who found nearly the same sequence: spleen, liver, GIT, pancreas, abdominal wall, genitourinary, adrenal, peritoneal, and biliary tract. In our study, splenic involvement was seen in 18 (12%) patients while it was involved in 20% of the patients according to Anis and Irshad.[16] GIT comes in order after spleen and bone marrow involvement. On the contrary, Gollub 2008[17] reported that GIT lymphoma is the most common location of the extranodal NHL (20%) but agreed with ours as the stomach was the mostly affected organ among GIT organs. Das et al.[11] and Manzella et al.[18] both stated that secondary liver involvement is more common than primary that is in concordant with our results. Different CT appearances were seen for osseous lymphoma with no pathognomonic sign in the form of normal CT, lytic, or sclerotic. The most common is lytic appearance. Furthermore, one case with lytic lesion appeared as a destructive bony lesion with no avidity on PET/CT after radiotherapy indicating fibrotic changes that agreed with Krishnan et al.[19] El-Haddad et al.[8] in their study stated that PET/CT is very effective in the detection of lesions in bone and bone marrow while CECT may be very effective in detecting recurrences in soft tissue, renal, and GIT extranodal lymphoma; however, in the current study, differences between CT and PET/CT were at bone, bone marrow, and splenic lesions, while in GIT lesions, they showed no significant difference. Paes et al.[1] in their study stated that sensitivity and specificity of PET/CT and CECT for the detection of extranodal sites are 88% and 50% for sensitivity, respectively, and 100% and 90% for specificity, respectively, while in our study, it showed for PET/CT 97% for sensitivity and 20% for specificity and CECT it showed 89.6% for sensitivity and 60% for specificity. This current study reported that assessment of extranodal involvement of lymphoma by PET/CT has 97% sensitivity, 20% specificity, 94.2% PPV, 33.3% NPV, and 91.7% accuracy for PET/CT; compared to 89.6% sensitivity, 60.0% specificity, 96.8% PPV, 30.0% NPV, and 87.5% accuracy for CT in relative agreement to Sollini et al.[20] whose results were as follows: sensitivity of 97%, a specificity of 92%, a PPV of 88%, an NPV of 98% and an accuracy of 94% recorded for PET/CT. In the study carried out by Fueger et,[21] they concluded that PET/CT was significantly better in detecting extranodal lesions than CT (P < 0.016) and PET (P = 0.06). Furthermore, Kamel et al.[22] who conducted their study on 37 patients, obtained similar results and supported this theory, as of 29 extranodal sites shown by standard reference, PET/CT was able to correctly determine 24 while PET and CT correctly detected only 15 and 16, respectively. The performance of PET/CT was significantly better in detecting extranodal lesions than PET (P = 0.016) and CT (P = 0.03). The same could be reached in this study as of 144 extranodal lesions shown by standard reference, PET/CT was able to correctly determine 130 while CT alone correctly detected only 120 (P < 0.009).

Conclusion

PET/CT is a multimodality technique that can accurately monitor the extranodal lymphoma, it can detect metabolically active lesions without CT structural changes and identify a viable tumor in normal size lymph nodes. PET/CT is effective than CT with enhanced contrast in evaluating extranodal lymphomatous infiltration, especially in spleen, bone, and bone marrow.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.