FDG-PET and Hodgkin Lymphoma: Are We Ready for a Functional Tumor Volume Definition?

Tumor Volume (TV) is an important independent prognostic factor in patients with Hodgkin Lymphoma (HL) and can be calculated manually using Contrast Enhanced Computed Tomography (CT) scan as proposed by Gobbi et al. (Gobbi PG et al, JCO, 2001; Gobbi PG et al, Cancer 2004). Definition of CT based TV (CT-TV) however is difficult to assess being time consuming and complex procedure.

Differently from CT, FDG-PET scan allows a better identification of tumor sites in HL and improves staging accuracy (Rigacci et al, Ann Hematol, 2007). Availability of digitalized images with PET makes it feasible to try to explore a PET based method to define TV (PET-TV).

We performed a pilot study to compare CT-TV and PET-TV on a series of phantoms and on a series of patients, to “optimize” the calculation of PET-TV in humans and to compare the new PET-TV with CT-TV. For the first part of the study 12 fillable objects characterized by different known volumes (VL) (range 0.5–1700 ml), shape and complexity were filled with a solution of water and FDG and then acquired with PET/CT. For the second part of the study 20 patients with HL, for whom both baseline CT and PET were available, were identified. Patients had a median age of 34 years, 6 had stage I, 11 had stage II (2 with Bulky and 2 with subdiaphragmatic disease), 3 had stage III or IV. For both parts of the study CT-VL and PET-VL were calculated separately and blindly by a radiologist and a nuclear medicine physician. CT-VL was defined manually on CT as originally reported. For definition of PET-VL a semiautomatic segmentation software (PET VCAR - Volume Computer Assisted Reading - GE) was used; volume contouring was defined automatically using different SUV thresholds (35-40-45-50-55-60%) or semiautomatically using an operator aided approach.

When used on phantoms both CT and PET allowed an accurate definition of VL with an overall difference with actual VL of +4% and +1%, respectively. The automatic approach for calculation of PET-VL was feasible with the best PET threshold for segmentation identified at 45%. Both CT and PET had the worst results with very small VL (<1.2 cm³) whose impact on global VL however is not significant. Dividing the VLs in “small”(range 0.5–26.5 cm³) and “big”(range 70–1700 cm³) CT-VL was –28% and +4% and PET–VL was +4% and +1%, respectively.

On patients CT-TV ranged from 13 to 840 ml. Regarding PET an automatic approach using 45% threshold defined with phantoms was first used but didn't allow sometimes to include all visible tumor; being the optimal threshold for patients variable an operator visual assessment evaluation was required.

Final PET-TV ranged from 29 to 536 ml. Overall using CT-TV as comparator, the use of PET allowed the identification of lower TV (median −11%: −47% to + 86%); in 12/18 PET-TV was lower than CT-TV with differences ranging from −9% to – 47%. In 6 patients PET-TV was greater than CT-TV (+1% to +86). In patient with lower PET-TV than CT-TV, this was mainly due to the presence of “hot” and “cold” areas within the tumor (especially big mediastinal masses) that cannot be read by CT and takes PET-TV closer to a “functional volume”(FV) rather than an anatomical volume only. In one case with the PET-TV higher than CT-TV, PET also included extranodal sites (bone and tonsils) that were not identifiable at CT. Finally TV calculation time/patient was 10 minutes (5 to 15) with PET, much less than with CT.

PET can be used to define VL in phantoms and TV in patients with HL. Although a fully automated method is feasible and effective in phantoms, a semiautomated approach must be used for PET-TV assessment in humans. With such approach PET-TV evaluation is feasible, easy and quick to assess. The use of PET-TV allows the identification of a new concept of “functional volume” (FV) that partially correlates with CT-TV but probably describes different aspects of the tumor. The clinical and prognostic relevance of FV in patients with HL must be further investigated.

No relevant conflicts of interest to declare.