Objectives: Surgical resection after preoperative chemotherapy in patients with non-small cell lung cancer might only be best for patients who are responders. We compared positron emission tomographic scanning with 2-fluoro-2-deoxy-d-glucose (FDP-PET scanning) with computed tomographic scanning to evaluate their ability to predict this response for the primary tumor, N1 and N2 lymph nodes.
Methods: All patients with non-small cell lung cancer who had an initial FDP-PET scan staging with tissue biopsy, neoadjuvant chemotherapy, repeat FDP-PET scanning, and repeat biopsies were prospectively studied.
Results: There were 34 patients (24 men; median age, 64 years). Eleven patients had N2 disease, and 7 had N1 disease. Twenty-seven patients received chemotherapy, and 7 patients received chemotherapy and radiation. All but 9 patients underwent resection. Statistical analysis showed FDP-PET scanning to be more specific (P <.0001), to have a higher positive predictive value (P =.0018), and to have a higher negative predictive value (P <.0001) than computed tomographic scanning for predicting residual tumor at the primary site. FDP-PET scanning was more sensitive (P <.0001) and more accurate (P <.0001), had a higher positive predictive value (P <.0001), and had a higher negative predictive value (P =.0002) than computed tomographic scanning for paratracheal nodes (number 2 and 4 lymph nodes). FDP-PET scanning had a higher positive predictive value (P <.0001) than computed tomographic scanning for the other N2 (numbers 5, 6, 7, 8, and 9) lymph nodes.
Conclusions: Repeat FDP-PET scanning is more specific and has a higher positive predictive value and negative predictive value than computed tomographic scanning for detecting residual tumor in the lung in patients with non-small cell lung cancer who have received preoperative chemotherapy. It is more sensitive and accurate for paratracheal N2 nodes as well. However, there is no significant difference in its detection of N1 lymph nodes.