Abstract Background: Processing multiple sentinel nodes (SN) is expensive and time consuming. The aim of this study was to analyze a series of SNs from node positive breast cancer patients in order to determine their diagnostic value and delineate a working algorithm.Materials and methods: A prospective database of 105 node positive breast cancer patients undergoing a SN biopsy was created. The SN biopsies were performed with Technetium sulfur colloid, blue dye (Isosulfan Blue or Patent Blue) or both. SN were defined as any blue, suspicious node or radioactive nodes up to 10% of the hottest node. Location of the nodes, criteria of SN definition (hot/blue/suspicious), ten second radioactive counts and pathologic status were recorded. A statistical analysis of the diagnostic value of those nodes was performed.Results: Three hundred and seventy five SNs were recorded in the database. The number of nodes per patient ranged from 1 to 15 with a median value of 3. Radioactivity was found in a total of 322 (86%) sentinel nodes and blue dye in a total of 105 (28%) nodes, with 92 (25%) nodes being both hot and blue. Fourty (11%) sentinel nodes were removed for suspicious appearance. One hundred fifty nine (42%) of the 375 sentinel nodes tested positive for disease. Among the 159 positive nodes, 139 (87%) were radioactive, 52 (33%) were blue, and 46 (29%) were both hot and blue. An analysis of the 322 radioactive nodes reveals that the most radioactive node was positive in 74/95 analyzable patients (77.9%). Consideration of the two most intense nodes was sufficient to diagnose nodal disease in an additional 15 patients, representing a significant increase in sensitivity to 93.7%, p<0.001. Examination of all other radioactive nodes diagnosed one more patient only but the increase in sensitivity, to 94.7%, was not significant (p=1.00).Sensitivity and Nodes Required for Diagnostic Tests Based on Hot Sentinel Nodes Nodes testedDiagnostic testSensitivity1 (95% CI2)Sensitivity1 (95% CI2)p-value3Median (range) per patientTotalAverage4Hottest node74/9577.9% (68.2-85.8%)-1951.0Two hottest nodes89/9593.7% (86.8-97.7%)<0.0012(1-2)701.9All hot nodes90/9594.7% (88.1-98.3%)1.003 (1 to 15)3033.21. Conditional sensitivity in SN-positive women. 2. 95% exact confidence interval. 3. Exact p-value, McNemar's test for a difference relative to the preceding row. 4. Average nodes required per patient.In patients with both radioactive and blue SNs, non radioactive blue nodes did not add significant diagnostic value.Non-radioactive blue nodes did not add significant diagnostic value. Nodes testedDiagnostic testSensitivity1 (95% CI2)Sensitivity1 (95% CI2)P-value3Median (range) per patientTotalAverage4Hot nodes596/10591.4% (84.4 to 96%)-33223.1Hot or blue nodes98/10593.3% (86.8 to 97.3%)0.5003 (0 to 15)3353.21. Conditional sensitivity in SN-positive women. 2. 95% exact confidence interval. 3. Exact p-value, McNemar's test for a difference relative to the preceding row. 4. Average nodes required per patient. 5. Nine false negatives are comprised of 4 patients with no hot nodes and 5 patients with negative hot nodes.Conclusions: Processing of the two hottest and the suspicious nodes is sufficient for the initial axillary staging. Additional SNs should be processed only in presence of nodal disease. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 1007.