Prognostic value of metabolic tumor volume on PET / CT in primary gastrointestinal diffuse large B cell lymphoma

Abstract

Primary gastrointestinal (PGI) diffuse large B cell lymphoma (DLBCL) is a relatively common disease. Recent studies indicate that measurement of maximum standardized uptake value (SUV_max) on pretreatment for ¹⁸F‐fluorodeoxyglucose PET is an important prognostic factor in PGI DLBCL. However, there is still an association between initial tumor burden and prognosis. Thus, in the present study, we investigated whether tumor volume by PET could have a potential prognostic value to predict the outcome. From 2006 to 2009, 165 Stage I E/II E PGI DLBCL patients were enrolled in the study. One hundred and five patients received cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab (R‐CHOP) only, whereas 60 patients underwent surgery plus R‐CHOP. Metabolic tumor volume (MTV) was defined initial tumor burden as target GI lesion above SUV, 2.5 by PET as a contouring border. Over a median follow‐up period of 36.6 months, receiver operating characteristic (ROC) analysis indicated that the best cut‐off values for MTV and SUV_max were 160.1 cm³ and 12.0, respectively. The estimated area under the ROC curve was higher for MTV than SUV_max. Thus, MTV was a better predictor for survival than SUV_max. In patients with a low MTV (<160.1 cm³), there were no significant differences in survival between patients undergoing R‐CHOP alone and surgery plus R‐CHOP (P = 0.347 for progression‐free survival [PFS]; P = 0.148 for overall survival [OS]). Conversely, in patients with a high MTV (>160.1 cm³), survival was longer in those who underwent surgery plus R‐CHOP than in those treated with R‐CHOP alone (P < 0.001 for PFS; P < 0.001 for OS). Multivariate analysis revealed that high MTV is an independent factor for predicting survival. Even in the era of rituximab, treatment of PGI DLBCL is not easy in patients with a high MTV. (Cancer Sci 2012; 103: 477–482)

The gastrointestinal (GI) tract is the most commonly involved extranodal site in primary non‐Hodgkin’s lymphoma (NHL), representing 10–15% of all cases of NHL and 30–40% of all extranodal sites.^{( 1 )} The most commonly involved site is the stomach (60–70%), followed by the small bowel, ileum, cecum, colon and rectum. Thus, diffuse large B cell lymphoma (DLBCL) is one of the most common primary GI (PGI) lymphomas.^{( 2} , ^{3 )} Optimal treatment of PGI DLBCL remains contentious and depends on the stage of the disease. For example, PGI DLBCL is treated with various therapeutic modalities, including surgery, chemotherapy and/or radiotherapy. However, many studies on PGI DLBCL were reported prior to the use of rituximab, with only a limited number published in the rituximab era.^{( 4} , ^{5 )}

Recently, several studies have been published supporting the usefulness of ¹⁸F‐fluorodeoxyglucose (FDG) PET in both the initial evaluation and subsequent follow‐up of PGI DLBCL.^{( 6} , ⁷ , ^{8 )} Abnormal FDG uptake measured by the maximum standardized uptake value (SUV_max) is correlated with cellular metabolism and can be influenced to some degree by the cell cycle speed. In fact, several reports have demonstrated the potential prognostic value of SUV_max on pretreatment ¹⁸F‐FDG PET in PGI lymphoma.^{( 9} , ^{10 )} However, the assessment of quantitative tumor burden remains a basic requirement for predicting prognosis and deciding on the most appropriate therapeutic strategy in NHL. Therefore, some investigators have proposed the measurement of PET volume, defined by absolute SUV as the contouring border, and have defined metabolic tumor volume (MTV) as the initial tumor burden.^{( 11} , ^{12 )} Comparison between these two factors (i.e. SUV_max and MTV) may redefine the active tumor burden state, using PET to predict the outcome, and may lead to the establishment of a comprehensive treatment strategy.

The aim of the present study was to determine whether tumor burden or disease activity measured by ¹⁸F‐FDG PET, as a new staging system, could have a potential role in predicting the outcome of two treatment strategies (i.e. surgery plus cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab [R‐CHOP] vs. R‐CHOP alone) in PGI DLBCL patients.

Materials and Methods

Patients.  From April 2006 to July 2009, 165 primary extranodal GI DLBCL patients with localized lymph node involvement were enrolled in the study and treated with R‐CHOP in seven medical centers (Pusan National University Hospital, Kyung‐pook National University Hospital, Seoul National University Bundang Hospital, Kosin University Gospel Hospital, Busan Paik Hospital, Gyeong‐sang National University Hospital, and Dong‐A University Medical Center). Study approval was granted by our institutional review board (Pusan National University Hospital) and written statements of informed consent were obtained from each patient.

Patients were included in the study if they had a Ann Arbor Stage IE or IIE primary extranodal localization, such as the GI tract, were available for clinical follow‐up, were deemed suitable on the basis of histological evaluation, and had a de novo DLBCL histotype. Patients were excluded from the study when adjacent organs or the liver were involved, if they presented with DLBCL secondary to low‐grade NHL or following other treatments, including radiotherapy and autologous stem cell transplantation.

Initial evaluation for diagnosis and PET staging.  Disease stage was evaluated according to the Ann Arbor staging system. All patients had undergone staging investigation, including ¹⁸F‐FDG‐PET, physical examination, blood and serum analysis, bone marrow (BM) aspiration and biopsy, and computed tomography (CT) scans of the neck, chest, abdomen, and pelvis. Pathologic diagnoses were preferentially made using upper and lower GI endoscopic biopsy. If symptoms were present, such as perforation, tendency for massive bleeding, or obstruction due to mass effect, diagnostic surgical resection was performed. In the small bowel site, surgical resection was performed in most patients when percutaneous needle biopsy was impossible to perform. Evaluation of the primary extranodal NHL was performed using the strict criteria of Dawson et al. ^{( 13 )}

Treatment schedule and response criteria.  Patients were treated every 3 weeks with six or eight cycles of R‐CHOP‐21 therapy, as suggested previously by Coiffier et al. ^{( 14 )} Pretreatment staging and response evaluation after six or eight cycles of R‐CHOP therapy were based on clinical examination, CT scans of the neck, chest, abdomen, and pelvis, BM biopsy, and PET. Responses were assessed according to revised International Workshop Criteria (IWC)^{( 15 )} and PET examinations were performed at the time of diagnosis and 3 weeks after completion of R‐CHOP therapy.

Measurement of MTV by ¹⁸F‐FDG PET/CT.  Dual‐modality PET/CT tomography was performed using a biograph (Siemens Medical Solution, Hoffman Estates, IL, USA), based on a dual‐slice helical CT and full‐ring PET. The FDG‐PET images were evaluated for regions of focally increased tracer uptake. In the target lesions of FDG tracer uptake, an SUV ≥ 2.5 as a contouring border was considered to represent lymphoma, as suggested by Freudenberg et al. ^{( 11 )} The CT images were used for correction of PET attenuation. Imaging reconstruction of corrected emission data was performed after Fourier transformation with AWOSEM algorithm (two iterations, eight subsets, 5‐mm Gaussian filter).^{( 16 )} Positive lymph nodes (LNs) were also corrected, and their volume was measured and added to the MTV of the GI site. The CT criterion for pathologic LN was that the LN exceeded 1.0 cm in all regions, except the groin. The PET image was also used to evaluate the area of focal tracer uptake; thus, an SUV ≥ 2.5 was considered as pathologic LN and the MTV was measured after correction of CT attenuation. The CT images were acquired at 130 mA, 130 kV, and a slice width (or 5 min and table feed) of 8 mm per rotation. Intravenous or oral contrast agents were used in all patients and a standardized breathing protocol was applied.

Assessment of ¹⁸F‐FDG uptake using SUV.  For all PET data, the tumor with the most intense FDG uptake among all foci was carefully identified, based on a graded color scale using red to indicate maximal counts. A volumetric region of interest (ROI) encompassing the entire tumor was drawn to ensure correct identification of the maximal count. Granulocyte–colony‐stimulating factor was stopped 48 h before PET and patients were instructed to fast for 6 h before injection of ¹⁸F‐FDG, except for glucose‐free oral hydration. Blood glucose levels were measured before injection of the tracer to ensure that levels were below 160 mg/dL.

Statistical analyses.  The Mann–Whitney U‐test was used to assess differences in the frequency of independent prognostic factors between Stage II and Stage III groups. Fisher’s exact test was used for descriptive statistical analyses of categorical data. The progression‐free survival (PFS) was calculated from the date of diagnosis to documented disease progression; observations were censored either on the date the patient was last known to be alive or on the date of death for patients dying as a result of causes unrelated to lymphoma or treatment. Overall survival (OS) was calculated from the date of diagnosis until either death as a result of any cause or the date last known to be alive. Both PFS and OS were estimated using the Kaplan–Meier method and differences were evaluated using the log‐rank test. Receiver operating characteristic (ROC) curves were constructed to estimate accuracy in predicting ideal cut‐off values for MTV and SUV_max. Estimations of sensitivity and specificity were based on the cut‐off values of MTV. Statistical analysis was performed using SPSS software for Macintosh version 15.0 (SPSS Inc., Chicago, IL, USA). P < 0.05 was considered significant.

Results

Table 1 summarizes the baseline characteristics of 165 PGI DLBCL patients. The median duration of follow‐up was 36.6 months, median patient age was 63 years (range 20–79 years), and there were 97 patients aged >60 years. Ninety‐three patients were male; 81 patients were in Stage IE and 84 patients were in Stage IIE. Forty‐two patients had elevated lactate dehydrogenase (LDH) at the time of diagnosis, whereas 29 patients had an Eastern Cooperative Oncology Group performance status (ECOG PS) > 2. Therefore, most patients (117 patients; 70.9%) were included in the low international prognostic index (IPI) group.

Table 1.

 Baseline characteristics of the patients

No. patients	165
Median (range) age (years)	63 (20–79)
No. men/women	93/72
Disease status
Age ≥ 60 years	97 (58.8)
Stage IE	81 (49.1)
Stage IIE	84 (50.9)
LDH > UNL	42 (25.5)
ECOG PS ≥ 2	29 (17.6)
B symptoms	24 (14.5)
Bulky disease	28 (17.0)
Median (range) MTD (cm)	5.6 (1.2–28.9)
Median (range) MTV (cm3)	132.5 (10.3–654.2)
Median (range) SUVmax	14.3 (2.5–71.0)
IPI score
Low (0–2)	117 (70.9)
High (3)	48 (29.1)
Extranodal sites involved
Stomach	85 (51.5)
Duodenum and jejunum	14 (8.5)
Distal ileum	41 (24.9)
Cecum	8 (4.8)
Ascending colon	11 (6.7)
Transversal colon	3 (1.8)
Descending colon	3 (1.8)
Treatment modality
Surgery followed by R‐CHOP	60 (36.4)
Gastric DLBCL	14 (8.5)
Intestinal DLBCL	46 (27.8)
R‐CHOP only	105 (63.6)
Gastric DLBCL	71 (43.0)
Intestinal DLBCL	34 (20.6)
Reason for surgery
Bleeding	15 (25.0)
Perforation	16 (26.7)
Obstruction	14 (23.3)
Mass resection due to deep site	15 (25.0)
3‐year survival for all patients (%)
Progression‐free survival	75.8%
Overall survival	80.0%
3‐year survival of patients with a low IPI score (%)
Progression‐free survival	81.1%
Overall survival	85.8%
3‐year survival of patients with a high IPI score (%)
Progression‐free survival	57.9%
Overall survival	60.5%

Unless indicated otherwise, data show the number of patients in each group, with percentages given in parentheses. DLBCL, diffuse large B cell lymphoma; ECOG PS, Eastern Cooperative Oncology Group performance status; IPI, international prognostic index; LDH, lactate dehydrogenase; MTD, maximum tumor diameter; MTV, metabolic tumor volume; R‐CHOP, cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab; SUV_max, standardized uptake value; UNL, upper normal limit.

Overall, 3‐year PFS and OS were 75.8% and 80.0%, respectively. In the low IPI group, the PFS and OS were 81.1% and 85.8%, respectively, compared with 57.9% and 60.5%, respectively, in the high IPI group.

Comparative baseline characteristics and outcomes according to treatment strategy.  Sixty patients (36.4%) underwent surgical resection followed by six to eight cycles of R‐CHOP therapy, whereas 105 patients (63.6%) received R‐CHOP therapy alone. Age, sex, disease status, and IPI scores were comparable between the two groups (Table 2). According to the revised IWC, the complete response rate was higher in the group of patients who underwent surgery followed by R‐CHOP therapy than in the group of patients undergoing R‐CHOP alone (P = 0.002). Conversely, the partial response rate was higher in the R‐CHOP only group (P = 0.002). Both stable and progressive disease states were comparable between the two groups (P = 0.633). Thus, the 3‐year PFS and OS were higher in the surgery plus R‐CHOP group than in the R‐CHOP alone group (PFS 88.3% vs. 68.6%, respectively, for PFS [P = 0.003]; 90.0% vs. 74.3%, respectively, for OS [P = 0.007]; Fig. 1a,b; Table 2).

Table 2.

 Comparison of baseline characteristics according to treatment strategy

	R‐CHOP only (n = 105)	Surgery plus R‐CHOP (n = 60)	P‐value
Age (years)	62 (20–79)	64 (22–79)	0.648
No. men/women	55/50	38/22	0.174
Disease status
Age ≥ 60 years (%)	60 (57.1)	37 (61.7)	0.571
Stage IE/IIE	49/56	32/28	0.411
WBC count (×109/L)	6320 (4120–10920)	6593 (4020–10600)	0.171
ALC (×109/L)	1599 (432–3630)	1568 (840–3010)	0.632
No. with LDH > UNL (%)	26 (24.8)	16 (26.7)	0.788
LDH (IU/L)	374 (132–2111)	380.5 (140–801)	0.566
No. with ECOG PS ≥ 2 (%)	19 (18.1)	10 (16.7)	0.970
No. with B symptoms (%)	14 (13.3)	10 (16.7)	0.560
No. with bulky disease	15 (14.3)	13 (21.6)	0.226
MTD (cm)	5.6 (1.2–28.9)	5.7 (1.3–14.0)	0.382
MTV (cm3)	132.1 (16.1–654.2)	149.2 (10.3–654.2)	0.230
SUVmax	13.7 (2.9–71.0)	15.4 (2.5–37.5)	0.192
IPI scores
Low (0–2)	81 (77.1)	36 (60.0)	0.944
High (3)	24 (22.9)	14 (40.0)
Response according to revised IWC
No. with CR (%)	65 (61.9)	51 (85.0)	0.002
No. with PR (%)	37 (35.2)	8 (13.3)	0.002
No. with SD and PD (%)	3 (2.9)	1 (1.7)	0.633
3‐year survival (%)
Progression‐free survival	68.6	88.3	0.003
Overall survival	74.3	90.0	0.007

Unless indicated otherwise, data are median values with the range given in parentheses. ALC, absolute lymphocyte count; CR, complete response; ECOG PS, Eastern Cooperative Oncology Group performance status; IPI, international prognostic index; IWC, international workshop criteria; LDH, lactate dehydrogenase; MTD, maximal tumor diameter; MTV, metabolic tumor volume; PD, progressive disease; PR, partial response; R‐CHOP, cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab; SD, stable disease; SUV_max, standardized uptake value; WBC, white blood cell count; UNL, upper normal limit.

Figure 1.

 (a,c) Progression‐free survival (PFS) and (b,d) overall survival (OS) in Stage IE and IIE patients according to treatment strategy (a,b) and metabolic tumor volume (b,d), as determined by PET/computed tomography. (a,b) Patients undergoing surgical resection followed by cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab (R‐CHOP) therapy had longer PFS and OS than patients receiving R‐CHOP therapy alone. (c,d) Similarly, patients with low metabolic tumor volume had longer PFS and OS than those with a high metabolic tumor volume.

Receiver–operating characteristic curve analysis of MTV and SUV_max.  In the present study, ROC curve analysis was used to calculate the accuracy of ideal cut‐off values to distinguish a low MTV group from a high MTV group, as well as a low SUV_max group from a high SUV_max group. The estimated area under the ROC curve (AUC_ROC) for MTV was 0.917, whereas the AUC_ROC for SUV_max was 0.696. Comparing the estimated AUC_ROC for MTV and SUV_max, we found that MTV was a better predictor of survival than SUV_max (Fig. 2). We used various cut‐off values for MTV and SUV_max to obtain a reasonable balance of sensitivity and specificity. Using 160.1 cm³ as the ideal cut‐off value for MTV resulted in a sensitivity of 90.9% and a specificity of 75.5%; for SUV_max, an ideal cut‐off value of 12.0 resulted in sensitivity and specificity of 90.9% and 42.4%, respectively.

Figure 2.

 Analysis of receiver operating characteristic (ROC) curves to determine whether metabolic tumor volume (MTV) or the maximum standardized uptake value (SUV_max) was the better predictor of survival in 165 patients with Stage IE and IIE primary gastrointestinal diffuse large B cell lymphoma (continuous variable). The area under the ROC curve (AUC_ROC) for MTV was 0.917 and the sensitivity and specificity of MTV were 90.9% and 75.5%, respectively. In contrast, the AUC_ROC for SUV_max was 0.696, with a sensitivity and specificity of 90.9% and 42.4%, respectively.

Survival outcome according to MTV cut‐off values and treatment strategy.  In the low MTV group (MTV < 160.1 cm³), 35 patients underwent surgical resection followed by R‐CHOP therapy, whereas 68 patients received R‐CHOP therapy only. Comparing outcomes according to treatment strategy in the low MTV group, PFS and OS were comparable between surgical resection plus R‐CHOP and R‐CHOP alone (P = 0.347 for PFS; P = 0.148 for OS; Fig. 3a,b). In the high MTV group (MTV ≥ 160.1 cm³), 25 patients underwent surgical resection followed by R‐CHOP therapy, whereas 37 patients received R‐CHOP therapy only. In the high MTV group, patients who underwent surgical resection plus R‐CHOP therapy had longer PFS and OS than patients receiving R‐CHOP therapy alone (P < 0.001 for PFS; P < 0.001 for OS; Fig. 3c,d). Of particular interest, patients receiving R‐CHOP therapy in the low MTV group had longer PFS and OS than those in the high MTV group undergoing surgical resection plus R‐CHOP therapy (P = 0.043 for PFS; P = 0.049 for OS).

Figure 3.

 (a) Progression‐free survival (PFS) and (b) overall survival (OS) in patients with low and high metabolic tumor volume (MTV). During the median follow‐up of 36.6 months, there were no significant difference in PFS (a) or OS (b) between patients receiving surgical resection followed by cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab (R‐CHOP) therapy and those receiving R‐CHOP therapy alone in the low MTV group (P = 0.347 and P = 0.148, respectively). However, the PFS and OS were shorter in the high MTV group receiving R‐CHOP therapy alone compared with all three other groups (P < 0.001 and P < 0.001, respectively). Interestingly, PFS and OS were longer in the low MTV group receiving R‐CHOP alone than the high MTV group undergoing surgical resection followed by R‐CHOP therapy (P = 0.043 and P = 0.049, respectively).

Clinical value of MTV compared with other prognostic factors.  To estimate the clinical value for predicting survival, we compared the clinical impact of high MTV and other potential prognostic factors, including age >60 years, complete response (CR) status, ECOG PS > 2, serum LDH levels over the upper limit of normal, high SUV_max (SUV_max ≥ 12.0), treatment strategy (e.g. surgical resection followed by R‐CHOP), bulky disease status (maximum tumor diameter [MTD] ≥ 10 cm), and high IPI score (IPI score = 3), using both univariate and multivariate analyses.

Univariate analysis revealed that all factors were significant for both PFS and OS (Table 3). In multivariate analysis, Cox’s proportional hazard analyses were performed to further investigate the prognostic value of high MTV, as well as other potential factors. According to multivariate analysis, a high IPI score (for PFS, hazard ratio [HR] = 3.286, 95% confidence interval [CI] = 1.657–6.517, P = 0.001; for OS, HR = 4.518, 95% CI = 2.157–9.463, P = 0.001; Table 3), high MTV (for PFS, HR = 15.725, 95% CI = 5.987–41.303, P < 0.001; for OS, HR = 16.257, 95% CI = 5.217–50.657, P < 0.001), and surgical resection followed by R‐CHOP therapy (for PFS, HR = 0.167, 95% CI = 0.072–0.387, P < 0.001; for OS, HR = 0.145, 95% CI = 0.057–0.369, P < 0.001) were independent prognostic factors for both PFS and OS, whereas other factors, including age >60 years, CR state, ECOG PS > 2, LDH > upper normal limit (UNL), high SUV_max, and bulky disease status were not significant in PGI DLBCL.

Table 3.

 Univariate and multivariate analyses for survival

	Univariate analysis			Multivariate analysis
	HR	95% CI	P‐value	HR	95% CI	P‐value
Progression‐free survival
Age > 60 years	2.742	1.304–5.766	0.004	2.255	0.948–5.361	0.066
CR state	0.990	0.503–1.946	0.035	0.597	0.264–1.270	0.173
ECOG PS ≥ 2	2.089	1.042–4.185	0.038	1.353	0.534–3.430	0.305
LDH > UNL	1.647	0.860–3.155	0.049	0.987	0.478–2.038	0.973
High IPI score	2.667	1.416–5.023	0.002	3.286	1.657–6.517	0.001
MTV ≥ 160.1 cm3	12.042	5.050–28.713	<0.001	15.725	5.987–41.303	<0.001
SUVmax ≥ 12.0	3.333	1.474–7.536	0.004	0.219	0.719–4.199	0.219
Surgery plus R‐CHOP	0.310	0.137–0.700	0.005	0.167	0.072–0.387	<0.001
Bulky disease (≥10 cm)	3.756	1.994–7.075	<0.001	0.851	0.417–1.735	0.657
Overall survival
Age > 60 years	2.937	1.273–6.774	0.012	2.586	0.932–7.179	0.068
CR state	1.186	0.551–2.552	0.036	0.627	0.257–1.529	0.305
ECOG PS ≥ 2	2.065	0.981–4.344	0.049	1.392	0.573–3.383	0.465
LDH > UNL	1.728	0.849–3.516	0.048	0.627	0.572–2.810	0.305
High IPI score	3.227	1.625–6.405	0.001	4.518	2.157–9.463	<0.001
MTV ≥ 160.1 cm3	13.037	4.582–37.088	<0.001	16.257	5.217–50.657	<0.001
SUVmax ≥ 12.0	4.896	1.720–13.931	0.003	2.761	0.906–8.416	0.074
Surgery plus R‐CHOP	0.316	0.130–0.767	0.011	0.145	0.057–0.369	<0.001
Bulky disease (≥10 cm)	3.300	1.640–6.641	0.001	0.828	0.431–1.959	0.828

CI, confidence interval; CR, complete response; ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; IPI, international prognostic index; LDH, lactate dehydrogenase; MTV, metabolic tumor volume; R‐CHOP, cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab; SUV_max, standardized uptake value; UNL, upper normal limit.

Discussion

In the present study, PGI DLBCL according to Ann Arbor staging presented as a localized disease, and the gastric area was the most commonly involved site. However, most of the gastric DLBCL patients (14 of 85 patients; 16.5%) in our series received R‐CHOP therapy only. Conversely, the most primary intestinal DLBCL patients (26 of 80 patients; 57.5%) underwent resection followed by R‐CHOP therapy. Primary gastric DLBCL has been treated with various modalities in the past, including surgery, chemotherapy, and radiotherapy, alone or in combination. Many previous studies have suggested that, particularly in Stage I and II patients, gastrectomy significantly improves survival.^{( 17} , ^{18 )} However, recently, Ferreri and Montalbán reported in their review that conservative non‐surgical treatment achieves equal or even better results than gastrectomy,^{( 19 )} although the role of consolidation radiotherapy is still debated. Bonet et al. ^{( 20 )} reported that four cycles of CHOP produced similar results to four cycles of CHOP plus irradiation of the field involved. Therefore, there was no advantage of combined treatment over chemotherapy alone in terms of survival. From these results, the authors concluded that chemotherapy was the decisive therapeutic component in gastric DLBCL. Interestingly, a recent retrospective study suggested that the clinical outcome in localized primary gastric DLBCL treated with three cycles of R‐CHOP plus radiotherapy tended to be similar to six cycles of R‐CHOP, with extremely favorable effects, whereas the prognosis of the advanced Lugano stage may be poor, even in the rituximab era.^{( 4 )} Conversely, in intestinal B cell lymphoma, previous prospective studies have demonstrated favorable outcomes for surgery/chemotherapy and another recent retrospective cohort study showed consistent results for surgical resection followed by chemotherapy in primary intestinal DLBCL.^{( 21} , ²² , ^{23 )} Thus, surgical resection followed by chemotherapy may be associated with excellent outcomes in localized primary intestinal DLBCL, but may be associated with unfavorable results in advanced disease, with patients with additional rituximab treatment having longer survival. In a comprehensive view of the above studies, the high‐risk patient group, including patients at an advanced stage, is still not easy to treat with a simplified strategy and a novel prognostic parameter is required in PGI DLBCL, even in the rituximab era.

It has been demonstrated that ¹⁸F‐FDG PET improves both primary staging and response assessment at completion of first‐line therapy for DLBCL.^{( 6} , ⁷ , ^{8 )} The present study in PGI DLBCL patients showed that the patient group undergoing surgery followed by R‐CHOP therapy had a higher response rate and longer survival than the group receiving R‐CHOP alone. Interestingly, in the low MTV group, there was no difference in survival between patients undergoing surgical resection plus R‐CHOP therapy and those treated with R‐CHOP alone. Moreover, surgical resection followed by R‐CHOP in the high MTV group showed a less favorable outcome than R‐CHOP therapy only in the low MTV group. The question remains as to whether surgical resection is needed in patients with lower tumor burden in PGI DLBCL and whether surgical resection followed by R‐CHOP therapy is sufficient to treat the high tumor burden in PGI DLCBL. This means that additional treatment strategies, such as radiotherapy as a consolidation therapy or more intensive chemotherapy, may be needed in patients with a high tumor burden.

In the present study, we compared the clinical of high MTV and high SUV_max as prognostic factors in PGI DLCBL using PET/CT. High MTV was a better predictor for survival than high SUV_max. Therefore, SUV_max may not reflect the true nature of the tumor. Furthermore, in the present study we compared the clinical relevance of several factors, such as high MTV, high SUV_max, high IPI score, treatment strategy (e.g. surgical resection followed by R‐CHOP), and bulky disease. We found that high MTV, high IPI score, and treatment strategy (resection followed by R‐CHOP) were independent prognostic factors for predicting survival. These results suggest that more intensive treatment strategies should be used in PGI DLBCL patients with high MTV and a high IPI score.

In the present study we calculated MTV using 2.5 as a cut‐off point for SUV according to the data reported by Freudenberg et al. ^{( 11 )} However, it is not clear whether the point would be an absolute numerical value and further well‐designed studies are needed to confirm our results. The present study has some limitations, such as a relatively short follow‐up time and non‐uniform diagnostic modalities used to assess the recurrence of disease.

In conclusion, a high IPI score remained a prognostic factor for predicting survival. Moreover, high MTV had important value in predicting outcome, as an indicator of both initial tumor burden and treatment strategy. Our results indicate that the importance of treatment strategy is weak in the low MTV group, whereas an intensive treatment approach could be needed in the high MTV group.

Disclosure Statement

The authors have no conflicts of interests to declare.