Secondary Logo

Share this article on:

Predictive Value of FDG PET/CT Versus Bone Marrow Biopsy in Pediatric Lymphoma

Badr, Salma, MSc*†; Kotb, Magdy, MD, PhD*; Elahmadawy, Mai Amr, MD, PhD*†; Moustafa, Hosna, MD, PhD

doi: 10.1097/RLU.0000000000002315
Original Articles

Purpose The aim of this study was to explore the positive predictive value and negative predictive value of FDG PET/CT. The prognostic impact of tumor burden of bone marrow infiltrates was diagnosed by FDG PET/CT at initial presentation.

Methods This retrospective study enrolled 140 pediatric patients with pathologically proven lymphoma (113 Hodgkin disease and 27 Non-Hodgkin lymphoma). All patients had pretherapy FDG PET/CT. Bone marrow biopsy (BMB), clinical, radiological, and follow-up data were also collected. The skeleton was divided into 8 segments, and a 5-point scoring system was used for assessment of BM infiltration burden.

Results Among the 140 lymphoma patients, FDG PET/CT revealed positive BM involvement in 41 patients; 2 of them were false-positive with negative BMB and regional MRI results. Positive predictive value was 95.1% for PET/CT compared with 100% with BMB. All patients diagnosed with positive BMI by BMB were detected by FDG PET/CT. On the contrary, BMB missed 25 patients (17.9%) with statistically significant difference. Negative predictive value was 100% for PET/CT compared with 80.2% for BMB (P < 0.05). FDG PET/CT upstaged 17.9% of the enrolled patients. Bone marrow involvement based on the 5-point scoring system was assessed. No significant difference was demonstrated in therapy outcome between patient with focal BMI (score 2) and extensive BMI (score 5; P = 0.06).

Conclusions FDG PET/CT has optimum negative predictive value compared with BMB in detection of bone marrow infiltrations in pediatric lymphoma with upstaging cases missed with BMB. Prognostic impact of BMI based on the 5-point scoring system reveals that the main influence is presence or absence of BMI rather than its tumor burden.

From the *Nuclear Medicine Unit, National Cancer Institute, Cairo University;

Children's Cancer Hospital; and

Nuclear Medicine Unit, Kasr Al-Ainy (NEMROCK Center), Cairo University, Cairo, Egypt.

Received for publication June 22, 2018; revision accepted September 5, 2018.

Conflicts of interest and sources of funding: none declared.

Authors' contributions: S.B. and M.A.E. conceived of the presented idea and wrote the manuscript with support from M.K.; M.K. helped in data interpretation and manuscript evaluation; H.M. participated in analysis and interpretation of data and supervised the findings of this work. All authors revised and gave final approval of the version to be submitted.

Availability of data and materials: Further data are available upon request. Please contact the corresponding author.

Correspondence to: Mai Amr Elahmadawy, MD, PhD, Nuclear Medicine Unit, National Cancer Institute, Cairo University, 1, Foam El-Khalig, El-Kasr Elaini St, Cairo 11796, Egypt. E-mail:

Lymphoma is a common pediatric malignancy. It has a tendency for intravascular infiltration within the bone marrow with an incidence ranging from 5% to 21% among Hodgkin disease (HD) patients and 30% to 50% among non-Hodgkin lymphoma (NHL) patients.1–4 Bone marrow involvement (BMI) is considered a systemic dissemination of the disease, which advances lymphoma patients to stage IV and adds significant morbidity and more frequent relapse that may necessitate intensified therapy protocol. Thus early detection and assessment of the extent of bone marrow infiltration as well as proper monitoring of therapy response is of clinical importance.5

Bone marrow examination can provide significant diagnostic and prognostic information in patients with different subtypes of lymphomas.6 It remains a cornerstone in the evaluation and monitoring of many lymphoma patients.7 Meanwhile, it is an invasive technique that entails a multistep process with technical challenges and diagnostic complexity that may result in false results and provoke misdiagnosis, especially if not integrated with morphology and immunohistochemical results.8,9

FDG PET/CT is an evolving diagnostic modality in the field of oncology. Its role in lymphoma is well established in different phases of disease assessment with high degree of accuracy particularly in FDG-avid lymphoma subtypes.10–12 Previous studies showed its reliable capability to mirror-image neoplastic bone marrow infiltrative process providing a noninvasive, whole-body mapping that copes with the known tendency of high-risk lymphomas to have a multifocal BMI.13,14

The role of FDG PET/CT in lymphoma is growing, and currently its use in lymphoma patients is widely accepted in different phases of disease with high degree of accuracy particularly in FDG-avid lymphoma subtypes. The use of FDG PET/CT in evaluation of BM infiltration in lymphoma patients is considered a hot focus for research.13–30 FDG PET/CT possesses the ability to noninvasively demonstrate multiple sites of BM and soft tissue involvement. It is known that lymphomas are tumors with much higher incidence of focal or multifocal BM involvement over diffuse infiltration,17,18 which has an adverse impact on histological evaluation of BM malignancies.19,20

In the current study, we aimed at investigating positive predictive value (PPV) and negative predictive value (NPV) of both FDG PET/CT scan and BM biopsy in the setting of initial assessment of pediatric lymphoma patients.

Back to Top | Article Outline


This retrospective study included 140 pediatric patients represented at Children's Cancer Hospital Egypt, between the periods of February 1, 2010, to December 31, 2015. All patients were referred to nuclear medicine departments at initial pretherapy phase. All patients had histopathologically proven lymphoma. Clinical information were extracted from the medical records including age, sex, methods of diagnosis, detailed pathology, imaging findings, response to treatment, and survival data.

Back to Top | Article Outline

Inclusion Criteria

The inclusion criteria include pediatric patients (age < 18 years), pathologically proved lymphoma at different stages of management, paired whole-body FDG PET/CT, and bone marrow biopsy (BMB) performed with time interval of 2 weeks.

Back to Top | Article Outline

Exclusion Criteria

The exclusion criteria include adults (age > 18 years), double primary, previous chemotherapy or radiotherapy, and recent surgical intervention for bone lesion.

All patients were informed about the details of the study. The ethical committee of NEMROCK Center and the radiation safety committee at the National Cancer Institute had given approval for study design.

Back to Top | Article Outline



The child was positioned on the procedure table (prone or lateral decubitus for posterior superior iliac crest and the iliac crest exposed). After ensuring that the child is appropriately immobilized and that the airway is not compromised by positioning, any topical ointment was removed and the skin was cleaned. The iliac crest was palpated, and the site of aspiration or biopsy was located. The puncture site was washed with antiseptic solution using a circular motion from the inner to the outer area, and then repeatedly washed 3 times. Sterile drapes were placed over the operative site. Anesthetizing the marrow site with 1 to 2 mL of local anesthetic (1% to 2% lidocaine without adrenaline) was done.

Back to Top | Article Outline


All BMBs are obtained from the posterior iliac crest. In most of the patients, the BMB was obtained bilaterally. The needle was held with the proximal end in the palm and the index finger against the shaft near the tip. With the stylet locked in place, the needle is introduced through the skin pointing toward the anterior iliac spine. Using gentle pressure, the needle was advanced with a slight twisting motion until it feels anchored to the bone; the stylet was removed, then using alternating clockwise and counterclockwise, the needle was rotated with 3 twists to the right and then to the left without advancing; repeated once again, then the needle was withdrawn using a rotary motion. Using a sterile gauze pad, manual pressure was applied to the site until the bleeding stops. The bone specimen was removed from the biopsy needle by introducing a probe through the distal end (this prevents specimen crushing). An adequate biopsy in children should contain at least 0.5 cm of well-preserved bone marrow.21 The specimen was dropped in formalin and labeled.

Back to Top | Article Outline

Interpretation of BMB

Trephine biopsy samples were decalcified and stained with hematoxyline and eosin (Gordon and Sweet's reticulin method). Accompanied marrow aspirates were stained with the May-Grunwald/Giemsa stain. All trephine biopsies were phenotyped with CD20 (B-cell marker), CD3 (T-cell marker), CD15, and CD30 (markers for Reed-Sternberg cells), as well as leukocyte common antigens (CD45). The marrow infiltration for lymphoma was interpreted by 2 pathologists who were blinded to the PET/CT results.

Back to Top | Article Outline


Patient Preparation

Parents were instructed that their children should fast for at least 4 to 6 hours before the study (to maintain low glucose and low insulin levels), but drink water to maintain good hydration except if sedation is indicated. The fasting blood glucose level was determined. The preferred fasting blood glucose is below 150 mg/dL.

Back to Top | Article Outline


FDG PET/CT study was done using a dedicated PET/CT scanner (Biograph, True-Point; Siemens). This camera integrates a PET scanner with a dual-section helical CT scanner (40 slice Emotion; Siemens) and allows the acquisition of coregistered CT and PET images in 1 session.

Scanning started 45 to 60 minutes after tracer injection of 5 to 10 MBq/kg, or 0.15 to 0.30 mCi/kg, with a minimum dose of 37 MBq (1 mCi). Intravenous contrast agent was administered in most patients. Initially, patients were examined in the supine position with arms elevated, and CT scanning was started with the following parameters: 400 mAs; 120 kV; slice thickness, 3 mm; and pitch, 1.5. The CT scans were acquired during normal respiration from skull vault reached caudally to the mid thighs. PET was performed immediately after acquisition of the CT images (5–7 bed positions; acquisition time, 2–3 min/bed position). The CT data were used for attenuation correction, and images were reconstructed as 3-mm slices applying a standard iterative algorithm (ordered-subset expectation maximization).

When necessary, sedation was used in accordance with guidelines before 18F-FDG PET/CT imaging to ensure patient immobilization and adequate image quality.

Back to Top | Article Outline


Images were interpreted at a workstation equipped with fusion software (advantage Window AW; Siemens) that provides multiplanar reformatted images and enables display of the PET images, CT images, and fused PET/CT images, which were interpreted by 2 experienced nuclear medicine physicians. The analysis was conducted on per-patient– and per-lesion–based analysis.

Back to Top | Article Outline


Qualitative (Visual) Assessment

For 18F-FDG PET/CT interpretation, any focal or patchy inhomogeneous uptake, superior-to-hepatic reference in the bone marrow was interpreted as abnormal FDG uptake; the CT images were revised for corresponding CT changes.

Back to Top | Article Outline

Quantitative Assessment

The SUVmax were recorded for the most active osseous lesion in each patient after manual application of the volumetric regions of interest on the transaxial attenuation-corrected PET slices, around the areas demonstrating the greatest accumulation of 18F-FDG and away from any nearby overlapping activity. Another sizable ROI was drawn over the normal liver where its SUVmax was considered reference activity.

Back to Top | Article Outline


Two experienced nuclear medicine physicians, blinded to the result of the BMB, read the 18F-FDG PET/CT scan and interpreted scans as positive or negative for BMI.

Back to Top | Article Outline

Positive PET/CT

PET/CT was interpreted positive for isolated/multiple focal uptake in the bone marrow more than the liver uptake and/or diffuse heterogeneous marrow involvement with sites of intense focal involvement with higher uptake than the liver.

Back to Top | Article Outline

Negative PET/CT

PET/CT was interpreted negative for BMI in the presence of diffuse homogenous marrow involvement with uptake less than or equal to liver.

FDG PET/CT findings were correlated with available pathological reports.

Back to Top | Article Outline

True-Positive PET/CT Results

PET/CT was interpreted as true-positive for agreed PET/CT and BMB in the detection of BMI, presence of morphologic changes on CT evaluation corresponding to the site of uptake in PET, disappearance of lesions on follow-up PET/CT after chemotherapy, and/or confirmation of these BM or osseous lesions using other radiological modalities (MRI).

Back to Top | Article Outline

False-Positive PET/CT Results

Patients with positive BMI in PET/CT study yet not meeting any of the previous criteria were considered as false-positive.

Back to Top | Article Outline


A negative PET/CT was considered as true-negative if both BMB and PET/CT were negative for BMI or no appreciable FDG uptake in bone marrow or bone in PET/CT and no associated morphological changes in CT.

Back to Top | Article Outline


PET/CT was considered as false-negative if BMB and/or other radiological (MRI) were positive and PET/CT was negative for BMI.

Back to Top | Article Outline


The assessment of therapy response was done depending on clinical, radiological, and Lugano criteria for assessment of FDG PET/CT (the new criteria for response assessment using CT and FDG PET/CT).22

Back to Top | Article Outline

Modified CT Response Criteria

Four categories for CT have been outlined: (a) complete radiologic response, all nodes less than or equal to 1.5 cm in longest diameter, disappearance of all CT findings of lymphoma; (b) partial remission, 50% or greater decrease in disease burden; (c) stable disease, less than 50% decrease in disease burden; and (d) progressive disease, new or increased adenopathy or new extranodal lymphoma.

Back to Top | Article Outline

PET Response Criteria

Response assessment with FDG PET/CT is based on metabolic activity, indicated by FDG uptake. The SUV serves as a marker of metabolic activity, and response assessment is now based on visual assessment of FDG uptake and categorized according to the “5-point scale.” The 5-point scale incorporates the Deauville criteria initially proposed for assessment on interim FDG PET/CT images. The 5-point scale includes the following categories:

  1. No FDG uptake > background.
  2. FDG uptake ≤ mediastinum.
  3. FDG uptake > mediastinum but ≤ liver.
  4. FDG uptake moderately > liver.
  5. FDG uptake markedly > liver and/or new lesions.

Note: New areas of FDG uptake are unlikely to be related to lymphoma.

The 5-point scale is applied to both interim and end-of-treatment FDG PET/CT response assessment. Four categories of response have been outlined as follows: (a) complete metabolic response—score of 1, 2, or 3; (b) partial metabolic response—score of 4 or 5 with reduced FDG uptake; (c) no metabolic response—score of 4 or 5 without significant change in FDG uptake; and (d) progressive metabolic disease—score of 4 or 5 with increased FDG uptake or with new lesions.

In interpreting the 5-point scale, a score of 1 or 2 is interpreted as negative for lymphoma, whereas a score of 4 or 5 is considered positive. A score of 3 likely also represents complete metabolic response at interim with resulting good prognosis and is therefore usually also considered as negative.

Back to Top | Article Outline


Relapse is the appearance of new FDG-avid lesions after attaining complete metabolic response.

Back to Top | Article Outline


Follow-up data for patients were retrieved from their medical records at the hospital where clinical and radiological data were obtained to evaluate patients' response to therapy till the last visit.

Interim, end-of-treatment, and follow-up PET/CT were performed for 138 at different time scales ranging from 2 to 15 months.

The follow-up data were used together with the pathological, other radiological modalities, and clinical data as a reference standard to differentiate between the false-positive results of PET/CT and false-negative results of BMB regarding BM infiltration.

Back to Top | Article Outline


Relapse (Event)-Free Survival

Relapse-free survival is a period between attainment of complete remission and occurrence of event (relapse).

Back to Top | Article Outline

Overall Survival

Overall survival is the period from beginning of treatment and death or the last follow-up.

Back to Top | Article Outline


Data were analyzed using SPSS win statistical package version 21 (SPSS Inc, Chicago, IL).

  • Numerical data were expressed as mean and standard deviation or median and range as appropriate.
  • Qualitative data were expressed as frequency and percentage.
  • χ2 Test (Fisher exact test) was used to examine the relation between qualitative variables. For quantitative data, comparison between 2 groups was done using either Student t test or Mann-Whitney test (nonparametric t test) as appropriate.
  • A P value of 0.05 or less was considered significant.
  • Receiver operator characteristic curve analysis used to find the best cutoff value for SUVmax to discriminate between progression and regression status as measure of prognosis with response, with the highest sensitivity and specificity.
  • The cutoff value for SUVmax was correlated with pathological types using χ2 test. A P value less than 0.05 was considered significant.
  • Kaplan-Meier method calculated all survival estimates. Other predictor and prognostic variables were related to survival using log-rank test. P value was set significant at 0.05 level.23
Back to Top | Article Outline


Clinicopathological Characteristics

This retrospective study was conducted on 140 pediatric patients with a biopsy-proven lymphoma. Patients were referred to the nuclear medicine unit at the Children's Cancer Hospital from February 1, 2010, to December 31, 2015. Their main clinicopathological characteristics are given in Table 1. There were 113 HD and 27 NHL patients. NHL patients were generally younger with a median age of 7 years (range, 2–16) compared with 9 years (range, 3–17) for HD. The sex ratio was similar in both groups with a male-to-female ratio around 3. A low frequency of B symptoms was noted with 29 of 113 HD and 4 of 27 NHL. Patients with HD were staged according to Ann Arbor staging system, whereas NHL patients were initially staged according to St Jude Children's Research Hospital (Murphy). A high frequency of stage II disease was noted on both categories (45% in HD and 33% of NHL). Stage IV was represented by 27 patient with HD having BM infiltration in 25 patients of them and 14 patients with NHL had BM infiltration revealed by clinicopathological, radiological data, and follow-up. The NS and MC subtypes represented the majority of patients with HD (94.7%), whereas the majority of patients with NHL belong to Burkitt subtype (55.6%; Table 1)



Back to Top | Article Outline
Bone Marrow Biopsy

All the patients underwent BMB from the posterior iliac crest for assessment of BM infiltration as a routine pretherapy staging workup with bilateral posterior iliac crest BMB in the majority of patients (131 patients representing 93.6%), whereas the remaining 9 patients (6.4%) underwent unilateral BMB. Only 14 instances of BM lymphomatous infiltration were approved in biopsy site. This infiltration is evident in 8 HD and 6 NHL patients.

Back to Top | Article Outline

Bone Marrow Infiltration in FDG PET/CT

Based on PET/CT results, positive bone marrow infiltration was seen in 41 patients (27 HD and 14 NHL).

Regional-based data analysis was used to evaluate the extent and burden of BM infiltration all over the skeleton using PET/CT. The skeleton was divided into 8 segments: spine (2), pelvis (2), skull (1), long bones (1), ribs (1), and scapula, clavicles, and sternum (1). The commonest site of involvement of bone marrow is seen in the pelvis and spine representing 27.3% and 26.3%, respectively, followed by long bones (15.3%); sternum, scapula, and clavicles (11.1%); and the skull bones (11.1%).

Accordingly, 5-point scoring system was used for assessment of BM infiltration burden:

  1. No BM infiltration.
  2. Any single lesion out of spine or pelvis.
  3. Two lesions outside (spine or pelvis) or single lesion in pelvis or spine.
  4. Two or more lesions including spinal or pelvic lesion or more than 2 lesions elsewhere.
  5. Extensive bone marrow infiltration.

Bone marrow involvement based on the aforementioned 5-point scoring system was assessed. The majority of the enrolled patients (70.7%) were scored 1 with negative BM infiltration, whereas the rest of patients had positive marrow infiltration (29.3%) and scored from 2 to 5 according to tumor load in the following descending pattern: score 5 (9.3%), score 4 (9%), score 2 (6%), and score 3 (5%).

The pattern analysis of bone marrow infiltrative lymphoma lesions in PET/CT showed focal marrow lymphomatous infiltrative lesions in 26 patients (63.4%) that was significantly higher than the diffuse pattern, which was presented in 15 patients (36.6%; P < 0.05). Moreover, associated morphological CT changes (mainly cortical) were seen in almost half of patient.

Back to Top | Article Outline

Clinicopathological and Follow-up Results

Clinicopathological and follow-up data using various laboratories and radiological imaging results were used to confirm presence or absence of BM infiltration as well as monitoring therapy response. Accordingly, BM infiltration was confirmed in 39 patients (14 biopsy proven, 19 combined PET and radiological CT changes, and the remaining 6 patients were proved during follow-up). On the other hand, 101 patients were free of bone marrow infiltration all through the duration of the study. A significantly higher frequency of positive bone marrow infiltration was demonstrated in NHL 14/27 patients (51.9%) compared with 25/113 patients (22.1%) in HD (P = 0.06; Fig. 1).



Further analysis of the distribution of positive bone marrow infiltration in different pathological types and subtypes of the included pediatric lymphoma patients was done. HD represented 64.1% of the patients with positive BMI, whereas NHL represented 35.9%. Furthermore, among different pathological subtypes of lymphoma; nodular sclerosis subtype was dominating in BM infiltration positive group (19 patients representing 48.7%) and followed by Burkitt lymphoma in 6 patients (15.4%; P = 0.02; Table 2).



Taking clinicopathological data as reference, results of PET/CT were evaluated, and according to pathological types, the patients divided into 2 groups:

  1. HD patients (n = 113): 27 patients had positive BM infiltration (25 true-positive and 2 false-positive results that were of nodular sclerosis sub type) and 86 patients were negative BM infiltration.
  2. NHL patients (n = 27): 14 patients had positive and 13 patients had negative results.

Significantly higher frequency for negative compared with positive marrow infiltration PET/CT scan results in HD patients. On the other hand, no significant difference between the number of positive and negative marrow infiltration in PET/CT results was demonstrated in NHL. No false-negative PET/CT scan results were demonstrated in both groups with high sensitivity indices (sensitivity and NPV) that amount to 100%. On the other hand, 2 false-positive instances with marrow infiltration PET/CT results were seen only in HD group. No false-positive PET/CT results were recorded in NHL group. Therefore, a nonsignificant decrease in specificity indices (specificity, PPV) and total accuracy were demonstrated in HD (97.7%, 92.6%, and 98.2%) compared with NHL patients, respectively (100%, 100%, and 100%.

Taking a follow-up and clinicopathological data as a reference standard, the bone marrow infiltration results of PET/CT and BMB were compared. The overall results show a high frequency for false-negative BMB results in respect to BM infiltration. In contrast, no false-negative results were reported for FDG PET/CT (Fig. 2). Therefore, the sensitivity and NPV were significantly higher for PET/CT amounting to 100% compared with BMB (35.89% and 80.16%, respectively). The same was not true in respect to specificity and PPV where a high comparable yield demonstrated for both techniques being 98% and 95.1%, and 100% and 100% for PET/CT and BMB, respectively. The aforementioned data were implemented on the overall accuracy that was significantly higher in FDG PET/CT compared with BMB (P < 0.05).



Back to Top | Article Outline

PET/CT Versus BMB in the Assessment of BM Infiltration in Different Pathological Types of Pediatric Lymphoma

Pathological-based data analysis was done in the attempt to compare PET/CT versus BMB in different pathological types of pediatric lymphoma. These comparisons were interpreted in the light of clinicopathological and follow-up data as a reference standard (Tables 3 and 4).





In respect to HD, a significantly wide difference in sensitivity and NPV between PET/CT (100% and 100%) and BMB (32% and 83.8%) are still existed. Also, high comparable yield for both modalities in respect to specificity indices is detected with no significant difference in between. The absence of false-negative results in PET/CT contributes to the aforementioned data. The same was not true in respect to BMB results in detection of marrow infiltration (Table 3).

In the assessment of BM infiltration in pediatric NHL patients, utmost high specificity indices were noticed for both techniques amounting to 100%. In contrast, significantly lower yield was obtained in BMB compared with PET/CT in respect to sensitivity indices in detection of bone marrow infiltration (P < 0.001; Table 4).

A comparison between mean values of SUVmax for lymphomatous bone marrow infiltrative lesions in HD and NHL lymphoma patients revealed no significant difference in the mean values of SUVmax in both groups (P = 0.13).

A comparison between FDG uptake in iliac crest (site of BM biopsy) PET/CT and BMB results in respect to BM infiltration was done in pediatric lymphoma patients. This comparison was done in the light of reference standard results. Biopsy site data analysis approved BM infiltration in 20 patients. Fourteen of the 20 were proved positive by BMB, whereas 18 of 20 showed increased FDG activity in FDG PET/CT. Disparity in results between both techniques was demonstrated in 8 patients. Two patients has negative FDG uptake at iliac crest that was positively infiltrated in BM biopsy results. In contrast, 6 patients had increased FDG uptake at iliac crest that was negative for BM infiltration by biopsy (Table 5).



Back to Top | Article Outline

Response and Survival Analysis

The difference in therapy outcome between pediatric lymphoma patient with positive and negative bone marrow infiltration is illustrated in Table 6. Complete remission was achieved in 90.7% of the enrolled pediatric lymphoma patients. There was a tendency for lymphoma patients with bone marrow to have a lower percentage (82%) of complete remission compared with those without bone marrow infiltration (94%) with a statistically significant difference (P = 0.05). Similarly, higher frequency of relapse and disease progression was demonstrated in positive marrow infiltrative pediatric lymphoma patients group (10% and 5%) compared with (4% and 2%) nonmarrow infiltrative patients, respectively (Table 6).



A follow-up period ranged from 6 to 65 months (mean, 17.14 ± 9.87 months) with 6 months' interval was used to assess the overall survival and relapse-free survival. The overall survival rate after the initial 6 months achieves 98.6% ± 0.010% patients that drops to 93.3% ± 0.028% after 2 years. Similarly, the recorded relapse-free survival rate at initial 6 months was 96.2% ± 0.017% that reduced to 91.9% ± 0.034% after 2 years.

No significant difference was demonstrated in overall survival rate at the initial 6 months and after 2 years in respect to presence or absence of bone marrow infiltration (P = 0.44). In the BM infiltration positive group at 6 months, the overall survival was 97.6% ± 0.024% that slightly reduced to 94.3% ± 0.04% after 2 years. Also in bone marrow infiltration negative group, no significant difference was recorded between 6 months (99% ± 0.01) and 2 years (92.9% ± 0.036) overall free survival (Fig. 3).



A significant difference was demonstrated in relapse-free survival rate in the initial 6 months and after 2 years in respect to presence or absence of bone marrow infiltration (P < 0.05). In the BM infiltration positive group, the 6 months' relapse-free survival was 90.9% ± 0.05% that significantly reduced to 81.8% ± 0.097% after 2 years. On the other hand, in bone marrow infiltration negative group, the 6 months' relapse-free survival was 98% ± 0.014% that slightly reduced to (95.4% ± 0.029%) at 2 years' follow-up with no significant statistical difference (Fig. 4).



Back to Top | Article Outline

Predictive Value of PET/CT in Terms of Disease Activity

Correlation between different mean values of SUVmax of the bone marrow infiltrate and therapy response in pediatric lymphoma patients with bone marrow infiltration was done. In respect to therapy response, the patients were divided into progressive (relapse and progression) and regressive (complete remission and partial remission) groups. No significant difference in the average value of the initial SUVmax could be detected among both groups (Table 7).



Back to Top | Article Outline

Receiver Operator Characteristic Curve Analysis for SUVmax Prognostic Cutoff Point

Receiver operator characteristic curve was used to mark a prognostic SUVmax cutoff point that discriminate between progressive and regressive course of disease after therapy in positively marrow infiltrated pediatric lymphoma patients with best compromise between sensitivity and specificity. Receiver operator characteristic curve marked SUVmax of 6.85 as a prognostic cutoff discriminator between both groups with a sensitivity of 66.7% and specificity of 75.8% (Fig. 5).



Back to Top | Article Outline

Predictive Value of PET/CT in Terms of Tumor Burden

Correlation between prognosis and BM tumor burden via the 5-point scoring analysis was done. Only 6 of the 101 patients with score 1 showed unfavorable prognosis. On the other hand, 6 of 39 with bone marrow infiltration had unfavorable prognosis, yet the magnitude of tumor burden in BM infiltrative group seemed to have no significant impact in respect to prognosis (Table 8).



Back to Top | Article Outline


In the current study, the overall frequency of bone marrow infiltration in pediatric lymphoma was 27.8%, which was more frequent among NHL representing (51.85%) compared with HD (22.1%). Similar data were reported by Cheng et al,1 who investigated the frequency of BM infiltration in 54 pediatric lymphoma patients. The overall incidence of bone marrow infiltration was 24%. Moreover, more frequent BM infiltration was demonstrated among NHL (9/23; ie, 39.1%) compared with HD (4/31; ie, 12.9%). Another study was done by Agrawal et al5 to assess the frequency of BM lymphomatous infiltration in 38 pediatric HD patients. BM infiltration was demonstrated in 8 patients (21.1%).

Currently, histological sample of the BM are usually obtained from the iliac crest and is considered to be the standard method for the evaluation of BM involvement in pediatric lymphoma. This method is based on the assumption that, in cases of BM involvement, tumor cells spread nonfocally through the whole marrow, which proved to be incorrect. Limitation of BMB includes invasive nature, sampling errors due to inhomogeneous random infiltration, and technical limitation (inadequate specimen, poorly decalcified/processed, inadequate sections, and inadequate stains). Moreover, accurate interpretation may necessitate sufficient experience to avoid common pitfalls and to provide conclusive results.9

Several published data in the literature showed a low yield for BMB in the assessment of BM infiltrates in lymphoma patients.24,25

The current work showed a low overall sensitivity (35.9%) for BMB in detection of BM infiltration in pediatric lymphoma. Moreover, a slightly better sensitivity was demonstrated in NHL (43%) compared with HD (32%). Inhomogeneous BM infiltration, technical insufficiency, and/or lack experience were contributed to such low yield.

Similar results were evident in a large study for 454 Hodgkin lymphoma patients done by El-Galaly et al25; BMB showed a low sensitivity (31%). Similarly, Berthet et al24 investigated BM infiltration in 133 newly diagnosed diffuse large B-cell lymphomas via BMB. The results revealed low sensitivity (24%) for BMB in detection of BM infiltration. Sampling error and/or technical limitation were the main contributing factors reported for low sensitivity in these studies. False-negative results may also increase when only BM aspirates and single-side BMB are performed. In fact, even bilateral BMB has significant false-negative findings due to the focal nature of BM involvement in patients with HD or NHL.20

In clinical practice, these high false-negative results of iliac crest BMB strongly enforce the need for a complementary noninvasive simpler tool for exploration of BM infiltration in lymphoma patients. This tool may entail exploration of the whole BM with high accuracy.

FDG PET/CT has the potential to outperform BMB in the evaluation of bone marrow infiltration in lymphoma.2,18

In the present study, 39 of 140 patients have positively infiltrated bone marrow. FDG PET/CT successfully identified those patients with marrow infiltration, and no false-negative FDG PET/CT results were obtained giving high sensitivity indices (ie, sensitivity and NPV) amounting to 100%. In contrast, BMB identify 14 of 39 bone marrow–infiltrated lymphoma patients with significantly lower sensitivity indices (sensitivity and NPV) of 35.9% and 80.2%, respectively.

On the other hand, no significant difference in specificity indices was demonstrated between FDG PET/CT and BMB in respect to lymphomatous BM infiltration in the current study. Only 2 false-positive instances with increase FDG BM uptake in FDG PET/CT were seen that may be attributed to noninfiltrative hypercellular marrow changes in HD patients. Therefore, comparable overall high specificity indices were demonstrated in FDG PET/CT and BMB in the current study. Both disparity and agreement between FDG PET/CT and BMB in respect to sensitivity and specificity indices, respectively, were achieved by many other authors in different pathological types of lymphoma. Several studies reported the superiority of sensitivity indices of FDG PET/CT; however, there is some controversy over safe replacement of FDG PET/CT to BMB in pediatric lymphoma patient. Also, Purz et al26 evaluated BMB and 18F-FDG PET/CT in the assessment of BMI in 175 pediatrics patients with HD stage above IIA. Significantly higher percent of BM infiltration were demonstrated by FDG PET/CT 26% compared with 4% by BMB in pediatric HD patients. Accordingly, they concluded that 18F-FDG/PET may safely substitute BMB in routine staging procedure in pediatric HD. Moreover, they found that PET/CT findings would have led to change in stage in 11% of patients and change in treatment in 5.7% of patients.

Similarly, higher sensitivity of FDG PET/CT compared with BMB in detection of BM infiltration in pediatric patients was also concluded by other authors.24,25 All show superiority of FDG PET/CT via detection of more number of BMI instances compared with BMB.

Cheng et al1 examined the role of 18F-FDG PET/CT in the initial evaluation of BMI in pediatric lymphoma patients including 31 HD and 23 NHL patients. PET/CT outperformed BMB by detecting 2 additional HD patients of the 31 that were not detected by BMB. Furthermore, in the study of Agrawal et al5 with 38 pediatric patients with HD, 18F-FDG PET/CT had sensitivity of 87.5%. 18F-FDG PET/CT was positive for BMI in 3/38 (8%) additional patients not detected by BMB.

Other studies investigated the role of FDG PET/CT in the detection of bone marrow infiltration in each pathological type of lymphoma separately; these studies showed that FDG PET/CT had significantly higher sensitivity and diagnostic accuracy in the detection of bone marrow infiltration as compared with BMB in both HD and NHL.24,25

Many studies reported that FDG PET was highly specific (specificity ranges from 91% to 100%) with 100% PPV for detection of BM involvement.3,4,27

In the current study, comparison between FDG uptake and BMB biopsy results at site of biopsy (iliac crest) was done. Biopsy site data analysis proved BM infiltration in 20 patients. Fourteen and 18 of 20 lymphoma patients with iliac crest marrow infiltration were positive by BMB and showed increased FDG activity in FDG PET/CT. FDG uptake was falsely negative in 2 patients at iliac crest biopsy site that was truly infiltrated as confirmed by BMB. In contrast, increased FDG uptake at iliac crest biopsy sites were noticed in 6 patients that had negative BMB. In those 6 patients, other sites for bone marrow infiltration were detected. Follow-up data showed normalization of FDG uptake at iliac crest biopsy site and other involved areas after specific therapy. Therefore, FDG uptake was considered truly positive in those patients. The 2 false-negative FDG PET/CT results can be explained on the basis of low metabolic activity or proliferative nature for that infiltration. On the other hand, the difference between increased FDG activity and BMB results is not well ascertained; however, technical errors and inadequate sampling in BMB may stand behind this discrepancy.

There are limited debatable data in the literature regarding the prognostic value of positive FDG PET/CT for BMI in pediatric lymphoma, especially those with negative BMB in respect to therapy response and outcome. Moreover, therapy-adapted protocols considering results of FDG PET/CT in evaluation of BM infiltration in pediatric lymphoma, particularly in those patients with negative BMB, are not achieved yet.

In the present work, the treatment response post first-line therapy was assessed in the light of presence or absence of BMI in the initial FDG PET/CT results. There was a significant difference in therapy response between negative and positive FDG PET/CT BMI; patients with negative FDG PET/CT for BMI shows better therapy response than those with positive FDG PET/CT for BMI (P = 0.05).

Similarly, Liang et al28 studied the prognostic value of BMI assessed by baseline PET/CT in 169 high-risk patients with diffuse large B-cell lymphoma. They found that worse survival outcome was found in patients with PET positive for bone marrow infiltration rather than those with negative PET.

In this work, both overall survival and relapse-free survival were assessed in pediatric lymphoma patients with positive and negative FDG PET/CT for BMI. No significant statistical difference in 2 years' overall survival were demonstrated between the patients with positive and negative FDG PET/CT bone marrow infiltration (94.3% and 92.9%, respectively; P = 0.44).

This result was matched with Hong et al29 study, which included 89 patients with diffuse large B-cell lymphoma treated with rituximab-CHOP; they found no significant differences in 2 years' overall survival in patients with positive or negative bone marrow infiltration findings on FDG PET/CT (59.4% vs 78.0%, respectively; P = 0.146).

Contrary to our results, Liang et al28 showed a significant difference in his study between PET/CT positive and PET/CT negative patients for BMI regarding overall survival. They stated that the 3 years' overall survival was higher in the patients with negative PET/CT for BMI than that with PET/CT BMI positive (84.2% ± 6.5% vs 44.1% ± 8.6%; P = 0.003). The aforementioned debatable results can be explained on the basis of the difference in the selected population and therapy protocol as well as the overall limited frequency of deaths in the initial 2 years after therapy in pediatric lymphoma patients.

Our results revealed significant statistical difference between FDG PET/CT positive and negative BMI results in respect to 2 years' relapse-free survival (81.8% ± 0.097% vs 95.4% ± 0.029%; P < 0.05).

Similar finding has been achieved by Berthet et al24 in NHL patients with DLBC type who concluded the independency of FDG PET/CT bone marrow status as a predictor of event-free survival.

In respect to HD, Purz et al26 made a comparison between a positive and negative BMI by PET in 175 HD patients in respect to event-free survival. They showed a trend toward a worse prognosis, if PET is positive. However, because of the overall low rate of events, the difference in the results is not statistically significant (P > 0.05).

Other authors supported the lower event-free survival in lymphoma with positive FDG PET/CT BM infiltration only in those patients confirmed by BMB compared with non–biopsy-proven ones in respect to event-free 2-year survival.30 On the contrary, Hong et al29 found no significant differences in event-free survival in patients with positive or negative BMI findings on FDG PET/CT.

Despite of the contradiction of the published data regarding the overall prognostic value of FDG PET/CT in respect to BMI in pediatric lymphoma patients, there is a more tendency toward unfavorable prognosis in PET/CT BMI positive patients; however, further large, more homogenous studies are needed to achieve solid conclusive results for validation in clinical practice.

PET/CT quantitative indices, especially SUVmax, may play a role as a prognostic indicator and therapy response monitoring in the assessment of lymphoma patients. The values of SUVmax may reflect the proliferative and metabolic activity of lymphoma lesions. Moreover, the values of SUVs before, during, and/or after therapy may help in therapy planning protocols in lymphoma patients. The higher values of SUVmax, the worse are the prognosis and vice versa. However, a cutoff point for the SUVmax values that discriminate between favorable and unfavorable prognosis in lymphoma patients is not achieved yet with few published data concerning this values.

In the current study, a value of 6.85 cutoff point for SUVmax for BMI lesions was demonstrated as a prognostic discriminator indices. The values of SUVmax higher than 6.8 have less favorable therapy response and vice versa. Close finding has been demonstrated by Liang et al.28 They demonstrated 8.6 as prognostic cutoff value for BMI lesions in FDG PET/CT in lymphoma patients. Among PET/CT BMI positive patients, patients with SUVmax of bone marrow infiltrate more than 8.6 were significantly associated with worse event-free survival and overall survival.

The extent of BMI may influence morbidity and mortality in many malignancies in pediatric patients. In the current work, 5-point scoring system was used for extent assessment of BMI and to test its prognostic value in pediatric lymphoma patients. In respect to the extent of BMI, a significantly higher frequency of focal BM infiltration was demonstrated in 63.4% of patients compared with diffuse BMI (36.6%) with significant difference (P < 0.05).

Moreover, no significant difference was demonstrated in therapy outcome between patient with focal and extensive BMI in the current work. This finding can be explained on the heterogeneity of selected population, variation in risk scoring and therapy regimen, as well as favorable outcome, especially with modern therapy protocols. On the other hand, BM infiltration may follow all or no rule in respect to prognosis in pediatric lymphoma patients; that is, single-focal BMI lesion may have similar outcome to extensive marrow infiltration. However, further larger studies with more homogenous group of patients may be required for clarifying such finding.

Back to Top | Article Outline

Strengths of the Study

  • The study includes a reasonable sample size of pediatric lymphoma patients.
  • It targets a topic with a relatively limited published data in the literature that entail the use of FDG PET/CT in the assessment of BMI in pediatric lymphoma patients.
  • FDG PET/CT quantitative analysis with determination of discriminator cutoff point was used for response monitoring and prognostic assessment.
  • Multiple sequential FDG PET/CT studies were used during follow-up to validate the current results.
Back to Top | Article Outline

Limitations of the Study

  • This is a retrospective study enrolled a heterogeneous group of patients population in respect to staging and pathological subtypes.
  • The results of current study were validated via clinicoradiological follow-up rather than target biopsy-proven ones, especially those outside the iliac bones. However, dynamic multiple biopsies are not suitable in clinical practice.
  • MRI bone marrow imaging was not performed in all patients.
  • Variation in therapy protocols data and duration of disease as well as follow-up period.
Back to Top | Article Outline


FDG PET/CT is a reliable diagnostic tool that provides whole BM mapping essential in the assessment of BMI in pediatric lymphoma patients. It has optimum NPV and appreciable role in upstaging cases missed by BMB. In clinical practice, FDG PET/CT should be considered in evaluation of BMI and upfronted to guide BMB and/or restrict its use for selected cases.

The prognostic impact of the FDG PET/CT–detected bone marrow infiltrates (in terms of therapy response) seems to be influenced by all or no rule (ie, presence or absence of BMI) rather than the tumoral burden of these infiltrates.

Back to Top | Article Outline


We recommend a prospective study using more homogenous group of patients in respect to pathological subtypes and risk factors in pediatric lymphoma patient with standardization of therapy protocol, duration of disease, and longer follow-up period. Also, we recommend comparing FDG PET/CT results in different clinical phases of the disease with other more objective diagnostic tool in the assessment of BMI in pediatric lymphoma patient, for example, MRI diffusion.

Back to Top | Article Outline


1. Cheng G, Chen W, Chamroonrat W, et al. Biopsy versus FDG PET/CT in the initial evaluation of bone marrow involvement in pediatric lymphoma patients. Eur J Nucl Med Mol Imaging. 2011;38:1469–1476.
2. Moulin-Romsee G, Hindie E, Cuenca X, et al. (18)F-FDG PET/CT bone/bone marrow findings in Hodgkin's lymphoma may circumvent the use of bone marrow trephine biopsy at diagnosis staging. Eur J Nucl Med Mol Imaging. 2010;37:1095–1105.
3. Schaefer NG, Strobel K, Taverna C, et al. Bone involvement in patients with lymphoma: the role of FDG-PET/CT. Eur J Nucl Med Mol Imaging. 2007;34:60–67.
4. Pelosi E, Penna D, Deandreis D, et al. FDG-PET in the detection of bone marrow disease in Hodgkin's disease and aggressive non-Hodgkin's lymphoma and its impact on clinical management. Q J Nucl Med Mol Imaging. 2008;52:9–16.
5. Agrawal K, Mittal BR, Bansal D, et al. Role of F-18 FDG PET/CT in assessing bone marrow involvement in pediatric Hodgkin's lymphoma. Ann Nucl Med. 2013;27:146–151.
6. Kwee TC, Kwee RM, Verdonck LF, et al. Magnetic resonance imaging for the detection of bone marrow involvement in malignant lymphoma. Br J Haematol. 2008;141:60–68.
7. Cheson BD. Staging and evaluation of the patient with lymphoma. Hematol Oncol Clin North Am. 2008;22:825–837.
8. Viswanatha D, Foucar K. Hodgkin and non-Hodgkin lymphoma involving bone marrow. Semin Diagn Pathol. 2003;20:196–210.
9. Wilkins BS. Pitfalls in bone marrow pathology: avoiding errors in bone marrow trephine biopsy diagnosis. J Clin Pathol. 2011;64:380–386.
10. Moog F, Bangerter M, Diederichs CG, et al. Lymphoma: role of whole-body 2-deoxy-2-[F-18]fluoro-D-glucose (FDG) PET in nodal staging. Radiology. 1997.
11. Lin EC, Alavi A. Lymphoma. In: PET and PET/CT: A Clinical Guide, 2nd ed. New York: Thieme Medical Publishers Inc; 2009.
12. Juweid ME, Stroobants S, Hoekstra OS, et al. Use of positron emission tomography for response assessment of lymphoma: consensus of the Imaging Subcommittee of International Harmonization Project in Lymphoma. J Clin Oncol. 2007;25:571–578.
13. Pakos EE, Fotopoulos AD, Ioannidis JP. 18F-FDG PET for evaluation of bone marrow infiltration in staging of lymphoma: a meta-analysis. J Nucl Med. 2005;46:958–963.
14. Chen-Liang T-H, Martin-Santos T, Jerez A, et al. The role of bone marrow biopsy and FDG-PET/CT in identifying bone marrow infiltration in the initial diagnosis of high grade non-Hodgkin B-cell lymphoma and Hodgkin lymphoma. accuracy in a multicenter series of 372 patients. Am J Hematol. 2015;90:686–690.
15. Salaun PY, Gastinne T, Bodet-Milin C, et al. Analysis of 18F-FDG PET diffuse bone marrow uptake and splenic uptake in staging of Hodgkin's lymphoma: a reflection of disease infiltration or just inflammation? Eur J Nucl Med Mol Imaging. 2009;36:1813–1821.
16. Muzahir S, Mian M, Munir I, et al. Clinical utility of 18F FDG-PET/CT in the detection of bone marrow disease in Hodgkin's lymphoma. Br J Radiol. 2012;85:490–496.
17. Arber DA, George TI. Bone marrow biopsy involvement by non-Hodgkin's lymphoma: frequency of lymphoma types, patterns, blood involvement, and discordance with other sites in 450 specimens. Am J Surg Pathol. 2005;29:1549–1557.
18. Ribrag V, Vanel D, Leboulleux S, et al. Prospective study of bone marrow infiltration in aggressive lymphoma by three independent methods: whole-body MRI, PET/CT and bone marrow biopsy. Eur J Radiol. 2008;66:325–331.
19. Moid F, DePalma L. Comparison of relative value of bone marrow aspirates and bone marrow trephine biopsies in the diagnosis of solid tumor metastasis and Hodgkin lymphoma: institutional experience and literature review. Arch Pathol Lab Med. 2005;129:497–501.
20. Wang J, Weiss LM, Chang KL, et al. Diagnostic utility of bilateral bone marrow examination: significance of morphologic and ancillary technique study in malignancy. Cancer. 2002;94:1522–1531.
21. Reid MM, Roald B. Adequacy of bone marrow trephine biopsy specimens in children. J Clin Pathol. 1996;49:226–229.
22. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2017;32:3059–3068.
23. Kleinbaum DG, Klein M. Kaplan-Meier survival curves and the log-rank test. Survival Analysis. New York, NY: Springer; 2012:55–96.
24. Berthet L, Cochet A, Kanoun S, et al. In newly diagnosed diffuse large B-cell lymphoma, determination of bone marrow involvement with 18F-FDG PET/CT provides better diagnostic performance and prognostic stratification than does biopsy. J Nucl Med. 2013;54:1244–1250.
25. El-Galaly TC, D'Amore F, Mylam KJ, et al. Routine bone marrow biopsy has little or no therapeutic consequence for positron emission tomography/computed tomography-staged treatment-naive patients with Hodgkin lymphoma. J Clin Oncol. 2012;30:4508–4514.
26. Purz S, Mauz-Körholz C, Körholz D, et al. [18F]Fluorodeoxyglucose positron emission tomography for detection of bone marrow involvement in children and adolescents with Hodgkin's lymphoma. J Clin Oncol. 2011;29:3523–3528.
27. Matthews R, Schuster M, Safaie E, et al. Fluorodeoxyglucose PET-CT findings following bone marrow harvesting. World J Nucl Med. 2012;11:19–21.
28. Liang JH, Sun J, Wang L, et al. Prognostic significance of bone marrow infiltration detected by PET-CT in newly diagnosed diffuse large B cell lymphoma. Oncotarget. 2016;7:19072–19080.
29. Hong J, Lee Y, Park Y, et al. Role of FDG-PET/CT in detecting lymphomatous bone marrow involvement in patients with newly diagnosed diffuse large B-cell lymphoma. Ann Hematol. 2012;91:687–695.
30. Khan AB, Barrington SF, Mikhaeel NG, et al. PET-CT staging of DLBCL accurately identifies and provides new insight into the clinical significance of bone marrow involvement. Blood. 2013;122:61–67.

lymphoma; pediatrics; FDG PET/CT; bone marrow biopsy

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.