Staging:
23/46 (50%) patients had pathologic stage I disease, eleven (24%) patients stage II disease, nine (19.5%) patients stage III disease (5x stage IIIa, 3x stage IIIb, 1x stage IIIc) and three (6.5%) patients were initially evaluated stage IVa disease.
Primary tumor staging: In the T1 group (total n=17, 37%) ten (22%) patients had a T1a LCNEC, five (11%) patients a T1b and two (4%) a T1c primary tumor. T2 LCNEC (total n=19, 41%) divided in 15 (33%) T2a and four (8.5%) T2b tumors. Finally, seven (15%) patients had a T3 tumor and three (6.5%) patients a T4 primary.
Lymph node involvement:
With respect to the N-staging 14 ± 7 lymph nodes (range 4 to 32 lymph nodes) were removed for sampling. The overall prevalence of a lymph node involvement was 32.6%. In detail, 31 (67.5%) patients were histopathologically staged pN0, eight (17.5%) patients pN1, five (11%) patients pN2 and two (4%) pN3.
Distant metastasis:
Three patients had proven distant metastasis. In an oligometastatic setting a female patient underwent resection of the LCNEC primary and a contralateral solitary pulmonary metastasis (fig. 3 G – J). At last follow-up seven months after resection she was free of disease. A second patient presented with a solitary brain metastasis. He received surgery for that brain metastasis and the pulmonary LCNEC primary in curative intent followed by chemo- and cerebral radiotherapy. The last patient presented with hepatic metastases and received mediastinoscopy for N-staging purposes.
FDG PET findings:
FDG PET/CT was performed 15.7 ± 12.7 days (range 1 to 60 days) before surgery or mediastinoscopy, respectively. The SUVmax of the primary tumor was 11.9 ± 7.1 (range 1.5 to 28.3) while the histologically verified lymph node metastases demonstrated a SUVmax of 9.7 ± 5.4 (range 3.5 to 19.7) (fig. 2). There was no significant difference between these uptake values (p = 0.29).
FDG PET/CT and histopathological evaluation:
When compared to the histopathological gold standard N-staging with FDG PET/CT was true positive in 14 patients, true negative in 27 patients, false positive in four patients and false negative in one patient (table 1). The false negative FDG PET/CT missed a single infiltrated lymph node (1/5 positive lymph nodes). This patient had neoadjuvant chemotherapy and the particular false negative lymph node demonstrated a low SUVmax of 2.0. However, in 2/4 of the false positive PET/CT scans a solitary lymph node demonstrated an increased FDG uptake while in the other 2/4 patients false positive tracer turnover was found in multiple mediastinal and hilar lymph nodes.
Taken together, this translates into a sensitivity of 93%, a specificity of 87%, an accuracy of 89%, a positive predictive value of 78% and a negative predictive value of 96% (table 1). Figure 3 depicts examples for a true positive and a false positive FDG PET/CT scan, respectively (fig.3).
Interrater variability:
Analysis of the interrater-reliability regarding the N-staging revealed a strong level of agreement with κ = 0.82 between both readers (p< 0.001) [12]. In eight patients the lymph node ratings of the readers were discordant.
PET/CT scanner subgroup analysis:
32 patients were investigated using the PET/CT scanner with analogue technique and 14 patients underwent the FDG scan in the fully digital PET/CT machine. Analyses of the FDG uptake in LCNEC primary tumors demonstrated a SUVmax of 12.1 ± 7.9 in the analogue PET/CT group and one of 11.3 ± 4.9 in the digital PET/CT group. For the true positive lymph node metastases SUVmax findings were 10.1 ± 5.5 in the analogue PET/CT group and 9.1 ± 5.6 in the digital PET/CT group, respectively.
Mann Whitney U-statistics revealed no significant differences between the two scanner types.
Calculations of sensitivity, specificity and accuracy resulted in 88%, 88% and 88%, respectively, in the analogue PET/CT group and in 100%, 88% and 93%, respectively, in the digital PET/CT group.
Pretreatment subgroup analysis:
37 patients directly underwent surgery as first therapy while nine patients received preoperative treatment with chemotherapy ± radiotherapy. In the surgery group the SUVmax of the LCNEC primary tumor was 11.8 ± 6.5 and that of the true positive lymph node metastases 10.3 ± 5.6. In the group with neoadjuvant treatment the SUVmax of the primary tumor was 12.4 ± 9.7. In those two patients with pretreated true positive lymph node metastases SUVmax were 6.7 and 5.5, respectively. Mann Whitney U-statistics revealed no significant differences between these subgroups.
Calculation of sensitivity, specificity and accuracy resulted in 100%, 85% and 89%, respectively, in the surgery first group and in 67%, 100% and 89%, respectively, in the chemo- ± radiotherapy pretreatment group.
Table 1: Results from histopathology and FDG PET/CT including sensitivity, specificity, positive and negative predictive values of FDG PET/CT
|
Histopathology
positive
|
Histopathology
negative
|
|
FDG PET/CT
positive
|
14
|
4
|
FDG PET/CT sensitivity: 93%
FDG PET/CT: positive predictive value 78%
|
FDG PET/CT
negative
|
1
|
27
|
FDG PET/CT specificity: 87%
FDG PET/CT: negative predictive value 96%
|