The prognostic value of 18F-FDG-PET metabolic heterogeneity in DLBCL after first-line immunochemotherapy

The majority of lymphomas, including diffuse large B-cell lymphoma (DLBCL), are ¹⁸fluoro-deoxyglucose (¹⁸F-FDG) avid. ¹⁸F-FDG is commonly used for staging of DLBCL, but its quantitative parameters, such as baseline SUVmax and total metabolic tumor volume (MTV), which has been demonstrated as prognostic factors, need defined robust cutoff values before they can be used in routine clinical practice.

A new parameter, ¹⁸F-FDG metabolic heterogeneity (also known as textural features), has recently emerged as a promising prognostic tool in solid tumors. However, its use in DLBCL has only been evaluated to a limited extent. Nicolas Aide and colleagues investigated the prognostic value of this new quantitative feature of ¹⁸F-FDG-positron emission tomography (PET) in patients with DLBCL treated with first-line immunochemotherapy using 2-year event-free survival (EFS). The results of the study were published in the journal European Radiology.¹

Methods

Data from patients with newly diagnosed with DLBCL were analyzed retrospectively.¹⁸F-FDG-PET studies were reconstructed using an algorithm with point spread function (PSF) modeling. For each patient, the total MTV was recorded and the volume of interest (VOI) structure of the largest target lesion (>10 cc) was used to determine the textural parameters. Data were then randomized to training and validation datasets, with 80% of patients attributed to the training dataset and remaining 20% of patients to the validation dataset. The training dataset was used to define the area under the curve (AUC) and optimal cutoff values for each metric from the 2-year EFS receiver operating characteristic (ROC) curves.

Patient and PET characteristics

• In total, data from 132 patients were analyzed
  • 89.4% of patients were treated with rituximab plus cyclophosphamide, doxorubicin hydrochloride, vincristine, and prednisolone
  • 10.6% of patients were treated with rituximab plus doxorubicin, cyclophosphamide, vindesine, bleomycin and prednisone
• The median age was 62 (20–81) years, and over half were female (53.8%)
• The median follow-up was 27 months
• The 2-year EFS was 77.3%
• The mean injected dose was 290 MBq
• There were no significant differences in clinical characteristics between patients who experienced an event at 2 years (2-year EFS 1) and those who did not experience an event (2-year EFS 0), except for age-adjusted international prognostic index (IPIaa) score and ¹⁸F-FDG uptake time (Table 1)

Table 1. Clinical and PET characteristics significantly different between patients according to 2-year EFS status

18F-FDG, 18fluoro-deoxy-glucose; IPIaa, age-adjusted international prognostic index; min, minute; SD, standard deviation *patients with no event at 2 years †patients who experienced an event at 2 years
Parameter	Whole population (n = 132)	2-year EFS 0* (n = 102)	2-year EFS 1† (n = 30)	p value
IPIaa, % 0–1 2–3	44.7 55.3	51.0 49.0	23.3 76.7	0.01
Mean 18F-FDG uptake time (SD), min	61 (5)	60 (5)	62 (4)	0.05

 Results

• Total MTV as well as certain second and third-order metrics were significant based on ROC analysis for 2-year EFS (Table 2)
• There were no significant differences in the metrics identified as significant by ROC analysis in patients with
  • Bulky vs non-bulky disease
  • Germinal center derived B cells (n = 52) vs activated B cells (n = 34) DLBCL
  • CD5⁺ vs CD5^-
• Total MTV, homogeneity, contrast, correlation, dissimilarity, long-zone emphasis (LZE), long-zone low grey-level emphasis (LZLGE), long-zone high-grey level emphasis (LZHGE), grey-level non-uniformity (GLNU) and zone percentage (ZP) were found by Cox univariable survival analysis to be associated with 2-year EFS
• LZHGE > 1,264,925.92 was found to be the only independent predictor of 2-year EFS by multivariate analysis (HR 2.84; 1.23–6.56; p = 0.01)
  • IPIaa 2–3 (HR 1.64; 95% CI, 0.63–4.27; p = 0.31) and MTV > 111.00 (HR 4.76; 95% CI, 0.57–39.41; p = 0.15) were not found to be predictive
• The predictive role of LZHGE on 2-year EFS was confirmed in the validation cohort (n = 27)
• 2-year EFS was 60.0% in patients with LZHGE > 1,264,925.92 vs 94.1% in patients with LZHGE ≤ 1,264,925.92 (HR 7.47; 95% CI, 0.83–66.99; p = 0.03)

Table 2. PET variables with significant difference based on ROC analysis for 2-year EFS

GLCM, grey level co-occurrence matrix; GLNU, grey-level non-uniformity for zone; GLSZM, grey-level size zone matrix; LZE, long-zone emphasis; LZLGE, long-zone low grey-level emphasis; LZHGE, long-zone high grey-level emphasis; MTV, metabolic tumor volume; ZP, zone percentage
Variable	Value	p value
Basic MTV total	> 111.00	0.01
GLCM Homogeneity Contrast Correlation Dissimilarity	> 0.55 < 3.76 > 0.60 < 1.39	0.04 0.04 0.03 0.04
GLDZM LZE LZLGE LZHGE GLNU ZP	> 24,829.55 > 1873.01 > 1,264,925.92 > 21.19 < 0.08	0.01 0.04 0.002 0.02 0.02

 Conclusions

The results of this study suggest that baseline ¹⁸F-FDG PET heterogeneity of the largest lymphoma lesion, in particular LZHGE, is a promising independent predictor of 2 year-EFS in patients with newly diagnosed DLBCL treated with immunochemotherapy. However, those findings need to be further explored and validated in a larger independent cohort of patients.