Imaging in Lymphoma: The Key Role of Fluorodeoxyglucose-Positron Emission Tomography

Abstract

This review focuses on the impact of new imaging guidelines for fluorodeoxyglucose-positron emission tomography (FDG-PET) on clinical practice and the future directions of lymphoma imaging.

Fluorodeoxyglucose-positron emission tomography (FDG-PET) has progressively changed lymphoma management over the past decade, and new imaging guidelines integrating FDG-PET for staging and response evaluation in lymphoma have been recently published [1, 2]. The present review focuses on the impact of these guidelines on clinical practice and the future directions of lymphoma imaging. In lymphoma, we are now facing a therapeutic dilemma, which explains why FDG-PET has gained wide acceptance. New treatments have improved the outcomes in the most common types of lymphoma; however, the classic prognostic factors fail to select the small percentage of patients with a high risk of relapse and treatment failure [3, 4]. In contrast, patients are also at risk of serious treatment-related morbidity and mortality; thus, overtreatment should be avoided in those patients who respond well to therapy. For these reasons, we need new prognostic and predictive factors, a precise determination of the initial disease extent, and an accurate and early assessment of the responsiveness to therapy. The objective is now to personalize treatment to improve cure rates in patients with adverse risk factors and to reduce toxicity in other patients without risking undertreatment or disease, respectively. Imaging using FDG-PET can satisfy many of these requirements. A case report of a patient with diffuse large B-cell lymphoma illustrates a typical use and value of FDG-PET in lymphoma management (Fig. 1).

Figure 1.

Case report: 60-year-old woman with a bulky mediastinal tumor. Eastern Cooperative Oncology Group classification, 1; age-adjusted international prognostic index, 2; lactate dehydrogenase greater than normal. Induction: rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisolone, with a 14-day schedule (six cycles). Positron emission tomography (PET) performed at baseline (PET0, top); after 2 cycles (PET2, middle), and after 4 cycles (PET4, bottom). PET corrected by attenuation, computed tomography (CT), and PET-CT-fused images are displayed for PET0, PET2, and PET4. PET2 showed a partial metabolic response (score 4, residual uptake in the mediastinal mass moderately increased compared with that of the liver). PET4 showed a complete metabolic response (score 2, some foci of residual uptake slightly increased compared with the mediastinal uptake). Consolidation therapy was autologous stem cell transplantation based on PET2 positivity in the context of a trial [34]. The patient was free of progression after 4 years.

Staging

Most lymphomas are FDG-avid, with the FDG avidity depending on lymphoma subtypes. The most common subtypes, including Hodgkin lymphoma (HL), diffuse large B-cell lymphoma (DLBCL), and follicular lymphoma are always avid at presentation. Mucosa-associated lymphoid tissue lymphomas, small lymphocytic lymphoma, extranodal marginal zone lymphoma, and cutaneous lymphomas are variably FDG-avid [5]. The FDG avidity is not only related to tumor cells but also to the proportion of environmental cells. In HL, in which a mass typically contains approximately 1% of tumor cells, the FDG uptake mainly results from accessory cells amplified by the cytokines produced by the Reed-Sternberg cells [6, 7]. In contrast, in DLBCL, which contains 90% of tumor cells, the FDG uptake mainly results from the tumor component. The degree of avidity expressed by the maximum standard uptake value (SUVmax) of the lymphoma sites is linked to lymphoma aggressiveness, and a SUVmax of 14 or more is suggestive of transformed lymphomas [8, 9].

PET-computed tomography (CT) is the most accurate staging technique in HL and FDG-avid non-HL (NHL), with an increased sensitivity for nodal and extranodal disease without the loss of specificity compared with CT alone (Fig. 2). Using PET-CT, the disease of 10%–30% of patients will be upstaged, many to a more advanced group, resulting in a more accurate definition of stage and prognostic scores and a change in a therapeutic consequence for about 15% of patients [10, 11]. Improving staging accuracy ensures that fewer patients will be undertreated or overtreated. Therefore, it is now recommended that PET-CT should be used for routine staging of FDG-avid nodal lymphomas as a reference standard and could be helpful for guiding biopsy (level of evidence 1) [1, 2]. A baseline PET-CT scan is also the best available tool for subsequent response assessment. Contrast-enhanced CT, when used at staging or restaging, should be performed as a “one-stop shop” in combination with PET-CT, if not already performed [12]. Low-dose unenhanced CT might suffice afterward, depending on the results of the baseline examination.

Figure 2.

Various patterns of fluorodeoxyglucose uptake in different subtypes of lymphomas: HL, DLBCL, FL, MZL, and AITL.

Abbreviations: AITL, angio-immunoblastic T-cell lymphoma; DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; HL, Hodgkin lymphoma; MZL, marginal zone lymphoma.

At present, PET-CT is not recommended routinely for lymphomas with low FDG-avidity (e.g., chronic lymphocytic leukemia/small lymphocytic lymphoma, extranodal marginal zone lymphoma, and some cutaneous and enteropathy-type T-cell lymphomas). However, in all cases when a transformation is clinically suspected, it is advised to perform PET-CT and to direct the biopsy to the lesion with the highest SUVmax value where technically feasible [9]. The extensive use of PET-CT in lymphoma has put the effectiveness of bone marrow biopsy (BMB) in question owing to the relative invasiveness of the latter procedure. Recently published data have suggested that focal FDG uptake in HL and aggressive NHL using PET-CT will accurately demonstrate bone marrow involvement [13–15]. Recommendations state that BMB is no longer indicated for HL (level of evidence 1) [2]. In DLBCL, baseline PET can detect bone marrow involvement in many patients who will not need confirmatory BMB. If the PET finding is negative in the bone marrow, a BMB can be considered if the physician believes it could change the prognosis or treatment [1]. In other lymphoma subtypes, PET is less accurate for detecting BM involvement. For example, in follicular lymphoma, PET misses more than 60% of bone marrow infiltration detected by BMB, which can be explained by the diffuse pattern and low grade of bone marrow involvement [16]. Baseline quantitative PET-CT has been explored as a prognosticator via the measurements of total metabolic tumor volume, which has the advantage of assessing the total tumor burden rather than just the presence or absence of a bulky mass [17]. Some limited results seem promising, showing that patients with the largest volumes have poor outcomes; however, large series are needed before these metrics can be used as a prognostic factor [18, 19].

The extensive use of PET-CT in lymphoma has put the effectiveness of bone marrow biopsy in question owing to the relative invasiveness of this procedure. Recently published data suggest that focal FDG uptake in HL and aggressive NHL using PET-CT will accurately demonstrate bone marrow involvement.

Response Evaluation and Outcome Prediction

The role of FDG-PET in the context of response evaluation and outcome prediction has recently been clarified, and consensus criteria for the use and reporting of interim (iPET) and end of treatment PET have been recommended [1, 2]. These criteria define FDG-PET as the standard of care for remission assessment (level of evidence 1) and introduce the concept of the metabolic response. Interim PET is usually performed during the first chemotherapy cycles (cycles 2–4). When performed after the first or second cycles, the result of the examination is a surrogate marker of chemosensitivity and could separate fast responding patients from nonresponding patients. Interim PET is aimed at assisting a risk-adapted and tailored therapy, leading to escalation or de-escalation [20].

The first large studies performed of aggressive NHL suggested that PET performed after two cycles could separate responders from nonresponders and that the result might be a good predictor of the prognosis [21]. The same observation was made in advanced Hodgkin lymphoma when PET was performed after two cycles of doxorubicin, bleomycin, vinblastine, dacarbazine (ABVD) with better positive predictive values and negative predictive values (NPVs) compared with NHL [22]. In both cases, iPET was reported as an independent predictor from the prognostic scores (international prognostic index and international prognostic score). Subsequent studies have reported major differences in the prognostic value of interim PET in DLBCL, as well as in HL. Actually, at this time, iPET was reported when the residual uptake was higher than a fixed reference background, which could be the nearby background, the mediastinal blood pool, or the liver, according to the studies. Consequently, for the same residual uptake, increasing the background turns a PET scan with positive findings to one with negative finding, which explains the conflicting results. The initial objective of the Deauville criteria introduced in 2009 was to define a common set of criteria for interim PET [23, 24]. The Deauville criteria are characterized by a 5-point scale grading the level of residual uptake according to different levels of background:

1. No uptake

2. Uptake less than or equal to that of mediastinum

3. Uptake greater than that of mediastinum but less than or equal to that of liver

4. Moderately increased uptake compared with liver

5. Markedly increased uptake compared with liver and/or new lesions (markedly increased uptake is taken to be uptake greater than 2–3 times the SUVmax in normal liver)

This scale is easy to use and improves interobserver reproducibility. Furthermore, the positivity threshold can be adapted in the case of escalating or de-escalating trials based on interim PET findings. The criteria were first validated for interim PET. A retrospective international validation study of 260 patients with advanced-stage HL has shown that iPET performed after 2 cycles and reported with a Deauville criteria (DC) cutoff between Deauville scores 3 and 4 for positivity was highly predictive of failure-free survival (FFS) (3-year FFS, 95% for PET-negative vs. 28% for PET-positive patients) [25]. A similar validation study of DLBCL has shown that iPET reported with the DC using the same cutoff was highly predictive of progression-free survival (PFS) (3-year PFS, 81% vs. 59%) [26]. Increasing the threshold to score 5 to define a positive scan reduced the false-positive rate (3-year PFS, 40% for PET-positive patients) but with poorer interobserver reproducibility. Recently, the DC have been used for reporting iPET findings after one cycle in patients with Hodgkin lymphoma, showing a high negative predictive value for iPET (NPV, 98%) [27]. However, at the moment, it is advised not to change treatment on the basis of the iPET results outside of a trial.

End of treatment PET is considered the best tool for the response evaluation in HL and DLBCL and has been recommended since 2007 by the International Harmonization Criteria [28]. It is now recommended to use the DC for end of treatment evaluation with the advantage of using a single method for both time points. Some studies have recently confirmed the usefulness of DC for end of treatment PET reporting. The Swiss observational study has shown in DLBCL treated with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisolone, with a 14-day schedule (R-CHOP14), that the end of treatment PET scan reported with DC with a cutoff for positivity between a score 3 and 4 was a good tool for remission assessment (2-year event-free survival, 70% vs. 27%) [29]. In high tumor burden follicular lymphoma, the French and Italian groups have shown in a pooled analysis that postinduction PET-CT reported with DC (score of ≥4 positive) was a good predictor of PFS and overall survival, confirming the usefulness of PET in follicular lymphoma [30]. The recommended Lugano classification for response assessment for interim or end of treatment PET using the Deauville criteria comprised different categories [2]: a score of 1–3 indicates a complete metabolic response (Fig. 3); a score of ≥4 indicates a partial metabolic response when uptake is reduced from baseline (Fig. 4), no metabolic response (NMR) if no change in uptake is seen compared with the baseline scan, and progressive metabolic disease (PMD) if increased uptake and/or new lesions are seen. NMR and PMD indicate treatment failure. The relatively high residual uptake that can be observed at interim and the end of treatment explains the level chosen for positivity. This has been shown for interim PET and confirmed for end of treatment PET owing to the generalized use of more aggressive therapy. The applicability of this classification requires that end of treatment PET must be performed a minimum of 3 weeks after chemotherapy (ideally, 6–8 weeks) and ≥3 months after radiotherapy and that interim PET must be performed as long as possible after the last chemotherapy administration and ideally just before the next chemotherapy cycle. Although the Lugano classification is easy to use, some limitations might exist when the level of the residual uptake is close to the uptake of the reference region, because the eye is sensitive to differences in contrast and not to differences in intensity [31]. Therefore, it is recommended to read the scan using a semiquantitative SUV scale, which allows one to “score” in a more objective/quantitative method the residual site in relation to the reference background (Fig. 5). In difficult cases, a further opinion should be obtained.

Figure 3.

Advanced-stage Hodgkin lymphoma. PET0 and PET2 after two cycles of escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone showing a complete metabolic response (score 1).

Abbreviations: PET0, baseline positron emission tomography scan; PET2, interim PET scan.

Figure 4.

Diffuse large B-cell lymphoma. PET0 and PET4 after 4 cycles of rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisolone, with a 14-day schedule. Partial metabolic response (score 4, mediastinal residual uptake higher than that of the liver but lower than the initial uptake).

Abbreviations: PET0, baseline positron emission tomography scan; PET4, PET scan after 4 chemotherapy cycles.

Figure 5.

Early-stage Hodgkin lymphoma. PET2 scan after two cycles of doxorubicin, bleomycin, vinblastine, and dacarbazine. The uptake in the left residual axillary lymph node seemed higher than the liver uptake because the contrast between the node and the axillary region is high (gray on white). Actually, the lymph node SUVmax was 2.9 and the liver SUVmax was 3.4.

Abbreviations: PET2, interim positron emission tomography scan; SUVmax, maximum standard uptake value.

The presence of a residual mass on the CT scan with negative PET findings observed at the end of treatment should be recorded. However, a German study has reported no difference in prognosis from a complete radiological response [32, 33]. CT remains very important for the response assessment in non-FDG avid lymphoma or in the absence of PET scan. In such cases, a CT-based anatomical response assessment is still recommended.

In refractory and relapsed HL and DLBCL, PET-CT is prognostic after salvage chemotherapy before high-dose (HD) chemotherapy and autologous stem cell transplantation (ASCT). Therefore, PET-CT can be used to identify poor prognosis patients before the decision is made to proceed with HD chemotherapy and ASCT. However PET-CT is discouraged for routine follow-up assessments, because it results in many false-positive studies with subsequent unnecessary investigations and the risk of increased patient anxiety.

In refractory and relapsed HL and DLBCL, PET-CT is prognostic after salvage chemotherapy before high-dose chemotherapy and autologous stem cell transplantation.

Conclusion

What will be the future of this technique in lymphoma? Quantitative PET is an active and fertile field of research. This includes the measurements of various metrics on baseline PET such as the metabolic tumor volume and tumor lesion glycolysis for better risk stratification and a more accurate response assessment at interim using the percentage of variation of uptake of the more active tumor (SUVmax) between the baseline and interim scans (ΔSUVmax) [34]. This parameter has proved superior to DC in evaluating the degree of response in DLBCL and HL and has been reported recently to be highly prognostic in a prospective trial, including 757 newly diagnosed patients having CD20-positive B-cell lymphomas (80% DLBCL) [35]. A new concept, “integrative PET”—a holistic approach combining these parameters with clinical, biological, or imaging data—has demonstrated in a short series of patients its capacity to improve patient risk stratification [20].

Author Contributions

Conception/Design: Michel Meignan, Martin Hutchings, Lawrence H. Schwartz

Provision of study material or patients: Michel Meignan, Martin Hutchings, Lawrence H. Schwartz

Collection and/or assembly of data: Michel Meignan, Martin Hutchings, Lawrence H. Schwartz

Data analysis and interpretation: Michel Meignan, Martin Hutchings, Lawrence H. Schwartz

Manuscript writing: Michel Meignan, Martin Hutchings, Lawrence H. Schwartz

Final approval of manuscript: Michel Meignan, Martin Hutchings, Lawrence H. Schwartz

Disclosures

Lawrence H. Schwartz: Icon Medical, Bioclinica, Novartis (SAB). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board