BACKGROUND This study evaluated the need for surveillance imaging in early-stage classic Hodgkin lymphoma (cHL) after planned combined-modality therapy (CMT). METHODS Primary early-stage cHL patients who underwent CMT were included. Positron emission tomography (PET)/computed tomography (CT), CT, or both were performed at the initial staging, during or after chemotherapy, and for at least 2 years during follow-up. Imaging studies and medical records were reviewed to determine if and when relapse had occurred. Radiation doses and costs were also calculated from follow-up imaging. RESULTS The study included 78 patients with a median follow-up of 46 months; 85% of the patients had stage II disease (32% with bulky disease). Four of 77 interim PET scans were positive; none of these patients relapsed during follow-up, which ranged from 24 to 80 months. After a total of 466 follow-up imaging studies (91% with CT and 9% with PET/CT), no cHL relapse was detected. Eleven abnormal findings were noted on surveillance imaging: 9 were false-positives, and 2 were second primary malignancies. The average cumulative dose per patient from follow-up imaging was 107 mSv, which translated into an estimated lifetime excess cancer risk of 0.5%; the estimated total costs were $296,817 according to Medicare reimbursements. CONCLUSIONS Surveillance imaging with either CT or PET/CT can be omitted safely for early-stage cHL treated with a combination of doxorubicin, bleomycin, vinblastine, and dacarbazine and radiation therapy because the risk of relapse is extremely low. This observation also applies to patients with bulky disease. The elimination of surveillance imaging will also reduce healthcare expenses and cumulative radiation doses in these predominantly young patients.

No abstract

5578 Background: Accurate assessment of cervical lymph nodes (LNs) for determination of metastatic involvement is of prime importance for management. Our objective was to determine the incremental value of combining metabolic and morphologic information in the differentiation between benign and malignant LNs. Methods: A total of101 pts with head and neck squamous cell (n=91) or thyroid cancer (n=6) or lymphoma (n=4) were included in the study. PET/CT and neck ceCT were acquired simultaneously or sequentially at staging (n=48) or restaging (n=53). In 131 LNs, variables evaluated includes SUVmax, size, shape (elliptical vs non-elliptical), presence of extracapsular extension (Ecext-irregular margins), necrosis, and fatty hilum. Histopathology (n=96) and 12 mo follow-up (n=35) were used for confirmation of the findings. ROC analyses determined the SUVmax cut-off. Results: Of131 LNs, malignancy wasconfirmed in 49 (37%).Results are shown in table. In the detection of malignancy, SUVmax of 4.5 yielded the best balance between the sensitivity and specificity, performing better than all CT variables alone. Combination analysis improved results only when SUVmax (4.5) was added to Ecext. However, best combination results were obtained at a SUV cut-off of 3.7. Higher SUV cut-off did not significantly improve overall performance of SUVmax alone (Table). Conclusions: In the differentiation of malignant from benign LNs, SUVmax (4.5) yields better results than CT variables alone. However combining a lower SUVmax (3.7) with Ecext produced the best results. Increasing SUVmax cut-off only produced a gain in specificity at a significant cost of sensitivity. [Table: see text]