FDG PET/CT Maximum Tumor Dissemination to Predict Recurrence in Patients with Diffuse Large B-Cell Lymphoma

Abstract

Purpose

We investigated the prognostic value of maximum tumor dissemination (Dmax), the distance between malignant lesions that were farthest apart, as assessed by fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT), and other clinical factors in patients with diffuse large B-cell lymphoma (DLBCL).

We investigated the prognostic value of maximum tumor dissemination (Dmax), the distance between malignant lesions that were farthest apart, as assessed by fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT), and other clinical factors in patients with diffuse large B-cell lymphoma (DLBCL).

Methods

Patients who underwent FDG PET/CT for initial staging and treatment response evaluation of DLBCL were reviewed retrospectively. Baseline Dmax, maximum standardized uptake value, total summation of all metabolic tumor volumes (tMTV), and total summation of all total lesion glycolysis (tTLG) were measured. The treatment response was evaluated at the interim and end of first-line treatment (EOT) using the Deauville score (DS). FDG PET/CT parameters and other clinical factors including sex, age, serum lactate dehydrogenase (LDH) level, stage, performance status, and the International Prognostic Index (IPI) were analyzed to identify factors prognostic of the time to progression (TTP) and disease-specific survival (DSS).

Results

A total of 63 patients were included. Univariate survival analysis identified Dmax (> 275 mm), tMTV (> 180 mL), tTLG (> 1300), interim DS (≥ 4), and EOT DS (≥ 4) as significant predictors of poor TTP. Serum LDH level (> 640 IU/L), IPI (≥ 4), tMTV (> 180 mL), tTLG (> 1300), interim DS (≥ 4), and EOT DS (≥ 4) were significant predictors of DSS. After multivariate survival analysis, Dmax (P = 0.008) and EOT DS (P = 0.005) were independent predictors of TTP. EOT DS was an independent predictor of DSS (P = 0.029).

Conclusions

Dmax at the time of diagnosis and the EOT response assessed by FDG PET/CT provide useful prognostic information additive to the IPI in patients with DLBCL.

Introduction

Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of non-Hodgkin lymphoma (NHL) worldwide and is frequently diagnosed at an advanced stage [1]. Although the introduction of intensive chemotherapy with the monoclonal antibody rituximab has improved outcomes for patients with DLBCL, about one-third of patients still do not achieve remission and experience relapsed disease [2].

The International Prognostic Index (IPI) includes 5 pretreatment characteristics, namely, age, serum lactate dehydrogenase (LDH) level, Ann Arbor stage, Eastern Cooperative Oncology Group performance status (ECOG PS), and extranodal involvement, and is used as a clinical tool to predict prognosis in patients with aggressive NHL [3]. Patients presenting with 4 or 5 negative prognostic factors are considered to be a high-risk group [4]. However, because the IPI was introduced at a time of chemotherapy-only regimens, a better predictor than the IPI is needed for patients with DLBCL treated with additional rituximab [5].

Fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT) is widely accepted as an efficient imaging modality for evaluating the metabolic activity as well as anatomical information of malignant lesions [6, 7]. Quantitative assessment of the metabolically active tumor burden using FDG PET/CT parameters such as the maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) has been proposed as a prognostic factor for various cancers, including lymphoma [8–12].

The 5-point scale Deauville score (DS) is a standard reporting tool for evaluating the response after treatment of lymphoma. The DS ranges from 1 for no uptake to 5 for marked uptake and the liver and mediastinum as the reference sites for comparison on FDG PET/CT [13]. FDG uptake significantly lower than the reference at the interim and the end of first-line treatment (EOT) is useful for assessing complete metabolic remission and predicting survival [14].

The maximum tumor dissemination (Dmax), the distance between malignant lesions that are farthest apart, assessed by FDG PET/CT, has been recently introduced as a prognostic factor in patients with DLBCL [15, 16]. The Dmax at the time of diagnosis can reflect the extent of tumor spread and may provide prognostic information for the outcome. Here, we investigated the prognostic value of FDG PET/CT parameters including Dmax and other clinical factors in patients with DLBCL.

Materials and Methods

Patients

Our institutional review board approved this retrospective study. Patient consent was waived because of the retrospective nature of the study and the analysis used anonymous clinical data. The medical records of consecutive patients with pathologically diagnosed DLBCL who underwent initial staging and treatment response evaluation at our center from June 2010 to October 2017 were reviewed.

The initial staging was performed according to the Ann Arbor staging and the Lugano classification criteria using contrast-enhanced CT and FDG PET/CT [7]. Bone marrow involvement of lymphoma was confirmed by bone marrow biopsy. Serum LDH level and ECOG PS before the treatment were assessed. The IPI, rather than the revised IPI (R-IPI) and the National Comprehensive Cancer Network IPI (NCCN-IPI), was also evaluated. Patients with a high risk of central nervous system (CNS) involvement underwent brain magnetic resonance imaging and cytological evaluation of cerebrospinal fluid. Patients with CNS involvement were excluded from this study.

Rituximab in combination with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) was given as the first-line treatment to all patients regardless of the disease stage. The treatment response after the third (interim) and sixth (EOT) cycle of R-CHOP was assessed using contrast-enhanced CT and FDG PET/CT according to the DS and Cheson criteria [17, 18]. Patients with a remnant lesion after the sixth cycle of R-CHOP considered in partial remission. Radiotherapy for the remnant tumor was given to elderly patients (older than 65), and high-dose chemotherapy with hematopoietic cell transplantation was given to younger patients. After the completion of the planned treatment, a follow-up evaluation was performed every 3 months for the first 2 years and every 6 months thereafter.

FDG PET/CT Imaging and Analysis

FDG PET/CT images were obtained using a GEMINI PET/CT scanner (Philips Medical Systems, Cleveland, OH, USA) from June 2010 to December 2011 for 14 patients and a GEMINI TF 64 PET/CT scanner (Philips Medical Systems, Cleveland, OH, USA) from January 2012 to October 2017 for 49 patients. Each PET/CT scanner was set for image acquisition and reconstruction as described earlier [8, 19]. Low-dose CT without contrast enhancement and PET emission images were acquired from the skull base to the mid-thigh in a supine position with/without the arms raised over the head. There was no lymphoma lesion outside of the axial field of view that we know of, such as on distal extremities.

Two experienced nuclear medicine physicians who were unaware of the clinicopathological results other than the presence of lymphoma reviewed the FDG PET/CT images to calculate FDG PET/CT parameters using the LIFEx package (version 7.0.1, Orsay, France) [20]. Any discrepancies were resolved by consensus. The PET images in DICOM format were imported into LIFEx. The SUV was defined as the concentration of FDG divided by the injected dose and normalized to the patient’s body weight. An SUV-based semiautomated contouring program was used to determine the volume of interest (VOI) of the lymphoma lesions. To define the tumor margin, the VOI comprised voxels with a threshold of 40% of the SUVmax in the tumor. If the VOI does not correctly mirror the visual interpretation of the FDG avid lymphoma lesion on the PET image, the SUVmax threshold was adjusted.

SUVmax, mean SUV, MTV, and TLG over the malignant lesions on attenuation-corrected PET images were calculated automatically by the program. TLG was calculated as the mean SUV multiplied by MTV. Total MTV (tMTV) and total TLG (tTLG) were summed for all MTVs and TLGs of all lymphomas in each patient, respectively. The farthest two lymphoma lesions were manually designated by the nuclear medicine physicians to calculate Dmax. The center of each malignant lesion was automatically defined from the 3-dimensional coordinates. The distance between the coordinates of the center of the 2 separate hypermetabolic lymphoma lesions was calculated using the Euclidean formula $D_{\mathit{ab}}=\sqrt{{\left(X_b,-,X_a\right)}^2+{(Y_b-Y_a)}^2+{(Z_b-Z_a)}^2}$.

Statistical Analysis

Receiver-operating characteristic curve analysis was performed to identify the cutoff thresholds for each characteristic of the patients with DLBCL for survival analysis. The durations of the time to progression (TTP) and disease-specific survival (DSS) from the date of diagnosis of DLBCL were evaluated using survival analysis. Deaths from causes unrelated to DLBCL were censored. Univariate analysis of prognostic factors and the significance of the difference between survival curves were tested using the Kaplan–Meier method and log-rank test. Multivariate survival analysis for independent prognostic factors was performed using the Cox proportional hazards model including the significant univariate variables. PASW Statistics for Windows (version 17.0; SPSS Inc., Chicago, IL, USA) was used. P values < 0.05 were considered to be significant.

Results

Patient Characteristics

The clinical characteristics of the 63 patients are summarized in Table 1. Most patients had an advanced stage of the disease (39/63, 62%). The Dmax was not significantly different between the patients with EOT DS 1, 2, 3 (329.4 ± 288.1 mm) and the patients with EOT DS 4, 5 (306.5 ± 292.5 mm, P = 0.794). Four patients (6%) were confirmed pathologically to have bone marrow involvement of lymphoma. Two of them showed descriptive malignant bone marrow lesions on FDG PET/CT, which could be measured as MTVs. Ten patients received an additional 2 cycles of R-CHOP after completion of the standard 6 cycles of chemotherapy. Four patients were treated with involved-field radiation therapy over the remnant suspicious lesion after R-CHOP. Twelve patients under 65 years of age underwent autologous peripheral blood stem cell transplantation (PBSCT). Three patients underwent additional chemotherapy before PBSCT (2 cycles of R-CHOP for 1 patient, 3 cycles of dexamethasone, L-asparaginase, ifosfamide, carboplatin, and etoposide for 2 patients).

Table 1.

Clinical characteristics of patients

Characteristic	Value
Sex (male/female)	28/35
Mean age (range)	57.3 ± 15.2 (21–87) years
Mean height (range)	162.2 ± 9.6 (142–187) cm
Mean weight (range)	61.2 ± 9.8 (37–87) kg
Ann Arbor stage (I/II/III/IV)	9/15/12/27
Mean serum LDHa (range)	909.5 ± 1311 (227–7928) IU/L
ECOG PSb (0/1/2/3/4)	24/32/6/1/0
IPIc score (0/1/2/3/4/5)	3/10/17/15/13/5
Mean Dmaxd (range)	318.5 ± 284.2 (0–815.0) mm
Mean Dmax according to the Ann Arbor stage (range)
Stage I	0.0 ± 0 mm
Stage II	138.0 ± 148.9 (36.0–592.5) mm
Stage III	586.3 ± 206.2 (120.6–757.8) mm
Stage IV	419.4 ± 262.3 (0–815.0) mm
Mean SUVmaxe (range)	20.4 ± 8.7 (3.0–42.9)
Mean tMTVf (range)	336.7 ± 657.0 (3–4706) mL
Mean tTLGg (range)	2783.1 ± 3991.2 (20–20,805)
Deauville score
Interim (1/2/3/4/5)	23/17/6/8/9
EOTh (1/2/3/4/5)	36/8/5/2/12

^aLDH: lactate dehydrogenase

^bECOG PS: Eastern Cooperative Oncology Group performance status

^cIPI: international prognostic index

^dDmax: maximum tumor dissemination

^eSUVmax: maximum standardized uptake value

^ftMTV: total metabolic tumor volume

^gtTLG: total total lesion glycolysis

^hEOT: end of first-line treatment

Survival Analysis

At the time of the last follow-up, 43 of 63 patients had survived after a minimum of 53.5 months of clinical follow-up (median 89.6 months). Seven of 63 patients died from causes other than DLBCL. All 7 patients showed no evidence of recurrence until the last follow-up. Two patients were dead due to old age (≥ 80). Two patients deceased from respiratory failure due to another underlying disease. One patient with the bed-ridden condition suffered sepsis due to a sore infection. Two patients’ social medical insurance was terminated without a written event on patients’ medical records. The DSS rate for 5 years was 78.3%. For all patients, the median TTP was 64.3 months (range 2.2–134.4 months) and the median DSS was 73.9 months (range 6.0–134.4 months).

In the survival analysis, the cutoff values were set for the categorization of age (> 60 years), stage (≥ III), serum LDH (> 640 IU/L), ECOG PS (≥ 2), IPI score (≥ 4), Dmax (> 275 mm), SUVmax (> 19), tMTV (> 180 mL), tTLG (≥ 1300), interim DS (≥ 4), and EOT DS (≥ 4). The significance of variables for predicting TTP and DSS in the univariate survival analysis is shown in Table 2. Univariate survival analysis showed that a high Dmax, high tMTV, high tTLG, high interim DS, and high EOT DS were significantly associated with poor TTP. High values for the serum LDH level, IPI score, tMTV, tTLG, interim DS, and EOT DS were significantly associated with poor DSS.

Table 2.

Univariate survival analyses of the time to progression and disease-specific survival

Variable		Time to progression			Disease-specific survival
(Number of patients)		HRa	95% CIb	P value	HR	95% CI	P value
Sex
Male	(28)	1.26	0.50–3.21	0.615	1.32	0.11–3.93	0.627
Female	(35)
Age, years
≤ 60	(31)	1.2	0.48–3.04	0.69	1.48	0.49–4.41	0.48
> 60	(32)
Ann Arbor stage
I–II	(24)	2.29	0.89–5.88	0.125	2.32	0.77–6.99	0.188
III–IV	(39)
Serum LDHc
≤ 640 IU/L	(38)	2.96	1.12–7.81	0.015	4.35	1.38–13.68	0.007
> 640 IU/L	(25)
ECOG PSd
0–1	(56)	1.06	0.24–4.8	0.13	0.83	0.13–5.5	0.859
2–4	(7)
IPI scoree
0–3	(51)	2.45	0.72–8.4	0.056	3.44	0.76–15.55	0.02
4–5	(12)
Dmaxf
≤ 275 mm	(33)	2.88	1.14–7.27	0.031	2.44	0.82–7.25	0.124
> 275 mm	(30)
SUVmaxg
≤ 19	(31)	1.75	0.69–4.41	0.232	2.33	0.79–6.91	0.147
> 19	(32)
tMTVh
≤ 180 mL	(36)	4.35	1.67–11.3	0.002	5.225	1.71–15.93	0.005
> 180 mL	(27)
tTLGi
≤ 1300	(36)	4.51	1.73–11.78	0.001	5.27	1.73–16.07	0.005
> 1300	(27)
Interim DSj
1–3	(56)	6.03	1.3–18.88	< 0.001	21.05	5.5–80.48	< 0.001
4–5	(17)
EOTk DS
1–3	(49)	9.52	2.53–38.81	< 0.001	18.68	4.1–85.1	< 0.001
4–5	(14)

^aHR: hazard ratio

^bCI: confidence interval

^cLDH: lactate dehydrogenase

^dECOG PS: Eastern Cooperative Oncology Group performance status

^eIPI: international prognostic index

^fDmax: maximum tumor dissemination

^gSUVmax: maximum standardized uptake value

^htMTV: total metabolic tumor volume

ⁱtTLG: total total lesion glycolysis

^jDS: Deauville score

^kEOT: end of first-line treatment

After multivariate survival analysis, Dmax and EOT DS were independent prognostic predictors of TTP. For DSS, only EOT DS was an independent prognostic predictor (Table 3). Kaplan–Meier curves of TTP and DSS for Dmax and EOT DS are shown in Fig. 1. Representative examples from patients with DLBCL with high and low Dmax at the time of diagnosis are shown in Fig. 2.

Table 3.

Multivariate survival analyses of the time to progression and disease-specific survival

Variable	Time to progression			Disease-specific survival
	HRa	95% CIb	P value	HR	95% CI	P value
Serum LDHc	0.4	0.11–1.52	0.178	1.46	0.3–7.22	0.641
IPI scored	–	–	–	2.2	0.55–8.89	0.268
Dmaxe	6.29	1.63–24.3	0.008	–	–	–
tMTVf	2.25	0.25–20.29	0.469	2.89	0.31–27.09	0.353
tTLGg	1.37	0.17–11.31	0.773	0.74	0.1–5.73	0.774
Interim DSh	1.06	0.16–7.09	0.955	4.92	0.75–32.16	0.096
EOTi DS	24.46	2.57–232.6	0.005	9.11	1.26–66.12	0.029

^aHR: hazard ratio

^bCI: confidence interval

^cLDH: lactate dehydrogenase s

^dIPI: international prognostic index

^eDmax: maximum distance between the two hypermetabolic malignant lesions

^ftMTV: total metabolic tumor volume

^gtTLG: total total lesion glycolysis

^hDS: Deauville score

ⁱEOT: end of first-line treatment

Fig. 1.

Kaplan–Meier analyses of the time to progression for (a) maximum tumor dissemination (Dmax), (b) end of first-line treatment (EOT) DS, and (c) overall survival for EOT DS

Fig. 2.

FDG PET/CT images in patients with diffuse large B-cell lymphoma (arrows) with previously known high-risk factors of relapse rather than Dmax (line). (a) A 38-year-old woman with Ann Arbor stage IV, ECOG PS 0, serum LDH 969 IU/L, IPI score 3, SUVmax 35.4, tMTV 362 mL, tTLG 8287, interim DS 4, EOT DS 4, and Dmax 253.1 mm. No recurrence was found until the last clinical follow-up at 107.7 months. (b) A 61-year-old man with Ann Arbor stage III, ECOG PS 1, serum LDH 461 IU/L, IPI score 3, SUVmax 11.4, tMTV 71.8 mL, tTLG 352.5, interim DS 2, EOT DS 1, and Dmax 743.9 mm. Recurrence was found after 16.6 months of follow-up. He survived until the last clinical follow-up at 77.3 months

Discussion

To the best of our knowledge, this is the first report to evaluate the prognostic value of the Dmax in patients with DLBCL in the Asian population. The present study found that Dmax at the time of diagnosis was independently associated with TTP and that EOT DS was independently associated with both TTP and DSS. Because patients with treatment failure after R-CHOP for DLBCL often have a poor prognosis, early recognition of high-risk patients with relapsed or refractory disease is necessary for stratifying the management of patients.

Although the IPI was developed almost 30 years ago, it remains widely used as a primary prognostic clinical scoring system in patients with DLBCL [3]. Since R-CHOP was accepted as the standard therapy for DLBCL and has improved outcomes, the R-IPI and the NCCN-IPI have been developed for identifying very high-risk patients with a 5-year overall survival < 50% [4, 21]. However, the IPI, R-IPI, and NCCN-IPI failed to identify the very high-risk patients who will have a survival clearly below 50% [5].

FDG PET/CT has been recently accepted as a significant complementary imaging modality to conventional CT for staging and evaluation of the treatment response in patients with DLBCL. Using FDG PET/CT to evaluate the baseline metabolic tumor burden has been suggested to help risk stratification [22]. Vercellino et al. reported that a high tMTV at the time of the diagnosis of DLBCL is a potential independent prognostic factor for survival [23]. Because it reflects the total volume of a metabolically active tumor, the tMTV can provide a more precise indicator of the tumor burden than previous markers such as serum LDH level. However, the assessment of tMTV using FDG PET/CT is time-consuming and difficult and requires an experienced operator. In this study, tMTV was a significant prognostic predictor of TTP and DSS in the univariate survival analysis, but it was not an independent prognostic predictor after the multivariate survival analysis. Because this study was performed through a retrospective review in a small number of patients, further prospective studies with more patients are needed to confirm the role of tMTV as a prognosticator in patients with DLBCL.

In most types of cancer, the metastatic spread of the disease is a strong reflection of tumor progression. In cases involving the malignant transformation of lymphocytes, which recirculate continuously between the blood and lymphatic systems, the absence of the conventional anatomic boundaries limiting the early spread of disease permits more rapid dissemination than in other cancers [24]. Thus, lymphoma dissemination is generally considered to reflect tumor progression less than in other cancers. However, as described in the Ann Arbor staging system, the dissemination of lymphoma has also been suggested to provide important prognostic information and to help guide therapeutic decisions. Until now, the quantitative assessment of entire lymphoma dissemination using conventional chest and abdomen CTs has been difficult. However, because FDG PET/CT acquires images from the patient’s head to thigh in a single view, the distance of the entire disease spread can be calculated easily using this quantitative method. As a result, Dmax is a simple and reproducible way to determine the tumor extent. In agreement with previous reports, Dmax was an independent predictor of TTP in this study [15, 16]. However, further investigation to find a better prognostic marker such as the number of organs involved or the number of nodal stations involved is necessary.

In 2007, the International Harmonization Project proposed the standardized use of PET for the assessment of the response of lymphoma [25]. The guideline was revised by the replacement of PET alone with PET/CT. At the first international workshop on PET in lymphoma in Deauville, France, in 2009, the DS was suggested as the standard reporting indicator [26]. The usefulness of applying FDG PET/CT to evaluate the response according to the DS has been investigated by several international studies [18]. The Lugano classification criteria recommend that FDG PET/CT should be used for response assessment in FDG avid lymphoma [7]. Previously, treatment modification according to the metabolic response evaluated using FDG PET/CT was suggested only for patients with limited-stage DLBCL [27]. However, a recent long-term report has suggested that patients with advanced-stage DLBCL who have a complete metabolic response on FDG PET/CT after at least 6 cycles of R-CHOP have a favorable outcome without additional radiation therapy [28]. Our study also shows that EOT DS may be an independent prognostic factor of both TTP and DSS.

Biological factors such as molecular subtype and genetic overexpression/rearrangement of DLBCL have been reported to correlate with outcomes. The activated B-cell-like subtype has an inferior outcome than the germinal center B-cell-like subtype [29]. In addition, DLBCL involving the overexpression of both MYC and BCL2 proteins or MYC and BCL2 and/or BCL6 rearrangements is associated with an adverse outcome after R-CHOP [30, 31]. However, further validation is required to integrate these into a clinical prognostic index.

Our study has several limitations. First and most notably, the results were obtained by a retrospective review of a small number of patients and a limited follow-up. This may have weakened the statistical power and introduced potential inherent bias. Further prospective studies with larger populations and longer follow-ups are needed to confirm these findings. Second, not all of the abnormally hypermetabolic lesions shown on FDG PET/CT were confirmed by histological diagnosis. Validation of lymphoma involvement by conventional imaging methods and serial follow-up may cause imprecise estimation of the entire disease extent. FDG non-avid lymphoma lesions may have been excluded or inflammatory lesions may have been included. However, because most cases of disease progression or death happened within 5 years and most of the patients who survived without these events were followed for > 5 years, it is unlikely that the conclusions of the current study would have changed significantly with a longer follow-up and histological diagnosis of all lymphoma involvement.

Conclusion

In patients with DLBCL, Dmax, as assessed by FDG PET/CT at the time of diagnosis, was an independent prognostic predictor of TTP. Complete metabolic remission after the end of first-line R-CHOP was an independent prognostic predictor of both TTP and DSS. Previously suggested factors such as tumor SUVmax, tMTV, tTLG, and other clinical parameters were not independent prognostic predictors in the present study. Thus, Dmax at the time of diagnosis and response evaluation at the EOT using FDG PET/CT may provide useful information about disease prognosis in patients with DLBCL.

Authors’ Contributions

J.J. collected data, performed data analysis, and wrote the manuscript. H.W.C. conceived the study, performed statistical analysis, provided supervision, and revised the manuscript. S.K. collected data and revised the manuscript. M.H.L. collected data and revised the manuscript. Y.S. performed data analysis, provided supervision, and revised the manuscript.

Availability of Data and Materials

Contact the corresponding author for data requests.

Declarations

Ethics Approval and Consent to Participate

This study was reviewed by the appropriate Ethics Committee and was, therefore, performed in accordance with the ethical standards laid down in the Declaration of Helsinki, revised in Brazil in 2013.

This retrospective study was waived from the need to obtain informed consent by our Institutional Review Board (KUMC2021-12–030).

Consent for Publication

All authors have read and agreed to the published version of the manuscript.

Conflict of Interest

Joon-Hyung Jo, Hyun Woo Chung, Sung-Yong Kim, Mark Hong Lee, and Young So declare no conflict of interest.