Introduction

Accurate documentation of both the anatomical extent of the disease and its response to therapy is crucial in the management of patients with lymphoma [1, 2]. Although computed tomography (CT) is the imaging modality of choice for staging of lymphoma, it is limited by the fact that the assessment of lymph node involvement is based on lymph node size and shape [3]. Especially post-therapeutic residual tissue may be difficult to differentiate from vital tumour [4] and thus present a diagnostic dilemma. [5].

Fluorine-18 fluorodeoxyglucose positron emission tomography (FDG-PET) is an established imaging technique for staging, restaging [6, 7, 8, 9, 10] and prediction of response to chemotherapy [11, 12, 13]. It has a higher sensitivity and specificity than anatomical imaging [6, 7, 8, 9, 10, 13, 14]. Furthermore, FDG-PET has a major impact on staging and managing lymphoma patients by changing the clinical stage in up to 44% and the clinical management in up to 62% of patients [14]. However, the low anatomical resolution of FDG-PET can be challenging, particularly in the abdomen. The combination of anatomical and metabolic imaging may overcome these shortcomings [15, 16, 17, 18].

The aim of our study was to determine the benefit of co-registration of FDG-PET and CT data in restaging of patients with lymphoma.

Materials and methods

Twenty-seven patients (16 men, 11 women, mean age 46 years, range 19–70)—18 with non-Hodgkin’s lymphoma (NHL) and 9 with Hodgkin’s lymphoma (HD)—who underwent clinical restaging post therapy were examined by whole-body FDG-PET/CT imaging.

Image evaluation was performed retrospectively. Written informed consent was obtained. The study was performed in accordance with guidelines issued by the local institutional review board.

FDG-PET/CT imaging

Dual-modality PET-CT tomography was performed on a biograph (Siemens Medical Solutions, Hoffman Estates, IL), which is based on a dual-slice helical CT and a full-ring PET tomograph. PET imaging was conducted 1 h after the administration of 360±20 MBq of FDG (3D mode, 5 min/bed position over five to eight bed positions). The CT images were used for PET attenuation correction [19]. Total examination time was 40 min or less. Image reconstruction of the corrected emission data was performed after Fourier rebinning (AWOSEM, 2 iterations, 8 subsets, 5 mm Gaussian filter).

CT images were acquired with 130 mAs, 130 kV, a slice width of 5 mm and a table feed of 8 mm per rotation. Intravenous and oral contrast agents were used in all patients [20]. A standardised breathing protocol was used [21].

Data analysis

All FDG-PET images were reviewed by two experienced nuclear medicine physicians in consensus, while CT images were read by two radiologists. The physicians were blinded to the other imaging modality. Following separate image evaluation, FDG-PET and CT data sets were read side-by-side in consensus. Finally, fused FDG-PET/CT data sets were viewed by the same physicians in consensus. Due to the higher sensitivity and specificity of FDG-PET, the absence of FDG uptake in an anatomical abnormality overrode the structural abnormality.

CT size criteria for individual lymph node groups were used to evaluate whether lymph nodes were pathological. Lymph nodes were considered pathological when they exceeded 1.0 cm in size in all regions except the groin, where they were considered pathological at diameters larger than 1.5 cm.

FDG-PET images were evaluated for regions of focally increased tracer uptake. In all lesions the maximum standardised uptake values (SUV) were determined [22, 23]. In an area of focal tracer uptake, a SUV of ≥2.5 was considered to represent malignancy.

For estimation of total lesion detectability, the imaging results were tabulated at lymph node regions: cervical, supraclavicular, paratracheal, mediastinal, hilar, axillary, coeliac, para-aortic, mesenteric, iliac and inguinal, which were further classified into four anatomical groups: (1) head and neck, (2) chest, (3) abdomen and (4) pelvis. Extranodal sites were classified separately. Each positive site within the same lymph node region was scored as one.

Standard of reference

When evaluating methods for tumour staging, a major problem is the definition of a “gold standard” with which the imaging modalities can be compared. Since in our study not all lesions could be evaluated by histology, we defined the sum of the imaging and follow-up data as the standard of reference for the extent and status of the disease. All patients were followed clinically for at least 12 months and all underwent follow-up CT (27/27). Furthermore, FDG-PET (n=13/27), ultrasound (n=23/27), and other imaging modalities (n=8/27) were acquired. Seven lesions were verified histologically. In patients with recurrent disease, response to subsequent therapy or tumour progression could be verified. Patients with absence of disease showed neither enlarging masses on follow-up CT and ultrasound nor pathological glucose metabolism on follow-up FDG-PET (9/13).

Statistical analysis

The consensus readings resulted in a set of data for each patient with respect to recurrence, lymph node and organ involvement. Sensitivity, specificity and positive and negative predictive values of the individual procedures were calculated from these data. The differences in the results were analysed statistically using a Fisher exact test. P values <0.05 were considered statistically significant.

Results

Of the 27 patients, 14 were diagnosed with recurrent lymphoma. Recurrence was located in 86 positive lymph nodes in 23/135 lymph-node regions. Extranodal involvement was found in two patients. Region-based sensitivity, specificity, positive predictive value, negative predictive value and accuracy are shown in Table 1. Differences between FDG-PET/CT and CT concerning sensitivity (P=0.005) and specificity (P=0.003) were statistically significant. Differences between FDG-PET and FDG-PET/CT were not statistically significant.

Table 1 Region-based (n=135) sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the different imaging modalities in lymphoma

FDG-PET/CT and FDG-PET and CT read side-by-side correctly characterised recurrent lymphoma in 13/14 patients, with the absence of disease in 13/13 patients. In one patient with histologically proven thyroid lymphoma, diagnosis was not possible with either modality. CT imaging alone was found to be correct in 67% (18/27) of patients (95% confidence interval 48–85%). FDG-PET imaging correctly characterised disease in 93% (25/27) of patients (95% confidence interval 81–100%). Patient-based sensitivity, specificity, positive predictive value, negative predictive value and accuracy of the different imaging modalities are summarised in Table 2.

Table 2 Patient-based (n=27) sensitivity, specificity, positive predictive value, negative predictive value and accuracy of the different imaging modalities in lymphoma

Based on the Ann Arbor staging system, tumour stage was correctly classified in 13 patients (48%) with CT, 20 patients (74%) with FDG-PET, 23 patients with FDG-PET and CT read side by side (85%), and 26 patients with FDG-PET/CT (96%). Differences between FDG-PET/CT and CT (P=0.002) were statistically significant, while no statistically significant differences were found between FDG-PET/CT and FDG-PET (P=0.1).

Compared with CT imaging alone, FDG-PET/CT correctly upstaged the disease in six patients (26%) and downstaged it in seven patients (26%). Table 3 lists the specifics of the up- and downstaging. Compared with FDG-PET imaging alone, FDG-PET/CT altered disease staging in 3/27 patients: the disease was upstaged in two patients (from I to III in patient 10, and from 0 to I in patient 13) and downstaged in one patient (from IV-L to IV in patient 7), as shown in Table 4. In two patients the treatment approach was changed based on FDG-PET/CT findings. One patient with recurrent HD presented a pathological retrocrural lymph node besides abdominal involvement characterising disease as stage III instead of I (patient 10). On FDG-PET alone this lymph node was localised incorrectly in the abdomen, while on CT it was missed owing to its small size (Figs. 1, 2). Another patient with a recurrent inguinal site of NHL was interpreted as having unspecific inflammatory tracer uptake on FDG-PET alone, as the lesion was not focal although it had a maximum SUV of 2.6 (patient 13).

Table 3 Data of patients with upstaging and downstaging, comparing co-registered FDG-PET/CT with CT alone
Table 4 Data of patients with upstaging and downstaging, comparing co-registered FDG-PET/CT with FDG-PET alone
Fig. 1A–C
figure 1

FDG-PET showed pathological tracer uptake (max. SUV 2.9) in a lymph node that was incorrectly localised in the abdomen on FDG-PET alone in the coronal (A), sagittal (B) and transverse (C) views (patient 10)

Fig. 2A, B
figure 2

Same patient as in Fig. 1. FDG-PET/CT (B) allowed exact anatomical localisation of a lymph node that was incorrectly localised to the abdomen on FDG-PET alone (Fig. 1) and was not pathologically enlarged on CT alone (A)

Discussion

Staging of patients with recurrent lymphoma has a great clinical impact on the choice of treatment regimen and survival. However, a major diagnostic challenge is posed by post-therapeutic non-viable residual masses, which can be demonstrated radiologically in up to 80% of patients with HD and in up to 40% of NHL patients [24, 25]. This is supported by our data showing a low specificity of CT imaging alone. FDG-PET was shown to be a more specific and sensitive imaging technique in this context. The overall sensitivities and specificities of FDG-PET are 87% and 92% respectively, compared with 93% and 10% for CT staging [6, 7, 8, 9, 10]. Our results obtained with FDG-PET imaging alone are in concordance with the literature, with an overall sensitivity and specificity of 86% and 100%, respectively. However, the low anatomical resolution of FDG-PET can be challenging. The aim of our study was to evaluate the benefit of co-registered FDG-PET/CT and to evaluate the impact on clinical management. Compared with side-by-side reading of FDG-PET and CT, combined FDG-PET/CT imaging changed the stage of disease and affected therapy in 2/14 patients (14%) with recurrent disease. Our data indicate that FDG-PET/CT further improves restaging in lymphoma. As patient survival depends on a stage-adapted therapy [26], the introduction of FDG-PET/CT imaging into the staging of lymphoma may improve patient therapy with an impact on patient survival. However, the current study did not address the issue of patient survival, and further studies are required to investigate this question.

In conclusion, compared with CT alone, FDG-PET/CT significantly improved sensitivity and specificity in restaging of lymphomas. Compared with FDG-PET alone, the number of correctly staged patients was increased, though not on the level of statistical significance. We conclude that FDG-PET/CT imaging is a promising technique to further improve restaging in lymphoma