Triage of patients with acute diverticulitis: are some inpatients candidates for outpatient treatment?

AIM: Current recommendations regarding the triage of patients with acute diverticulitis for inpatient or outpatient treatment are vague. We hypothesized that a significant number of patients treated as an inpatient could be managed as an outpatient. METHOD: A retrospective cohort study was carried out of 639 patients admitted for a first episode of diverticulitis. The diagnosis of acute diverticulitis was confirmed by computed tomography (CT). The endpoints included length of stay, need for surgery, percutaneous drainage and mortality. Patients were considered to have had a minimal hospitalization, defined as survival to discharge without needing a procedure, hospitalization of ≤ 3 days and no readmission for diverticulitis within 30 days after discharge. RESULTS: Of 639 patients, 368 (57.6%) had a minimal hospitalization. Female gender and CT scan findings of free air/fluid were negatively associated with the likelihood of minimal hospitalization. The presence of an abscess < 3 cm and stranding on CT did not predict the need for a higher level of care. Despite the statistical significance of several patient-level predictors, the model did not identify patients likely to need only minimal hospitalization. CONCLUSION: Most patients admitted with acute diverticulitis are discharged after minimal hospitalization. Free air/liquid in a patient admitted for acute diverticulitis indicates a more severe clinical course.

Does complete resection of melanoma metastatic to solid intra-abdominal organs improve survival?

BACKGROUND: Patients with distant melanoma metastases have median survivals of 4 to 8 months. Previous studies have demonstrated improved survival after complete resection of pulmonary and hollow viscus gastrointestinal metastases. We hypothesized that patients with metastatic disease to intra-abdominal solid organs might also benefit from complete surgical resection. METHODS: A prospectively acquired database identified patients treated for melanoma metastatic to the liver, pancreas, spleen, adrenal glands, or a combination of these from 1971 to 2000. The primary intervention was complete or incomplete surgical resection of intra-abdominal solid-organ metastases, and the main outcome measure was postoperative overall survival (OS). Disease-free survival (DFS) was a secondary outcome measure. RESULTS: Sixty patients underwent adrenalectomy, hepatectomy, splenectomy, or pancreatectomy. Median OS was significantly improved after complete versus incomplete resections, but median OS after complete resection was not significantly different for single-site versus synchronous multisite metastases. The 5-year survival in the group after complete resection was 24%, whereas in the incomplete resection group, there were no 5-year survivors. Median DFS after complete resection was 15 months. Of note, the 2-year DFS after complete resection was 53% for synchronous multi-site metastases versus 26% for single-site metastases. CONCLUSIONS: In highly selected patients with melanoma metastatic to intra-abdominal solid organs, aggressive attempts at complete surgical resection may improve OS. It is important that the number of metastatic sites does not seem to affect the OS after complete resection.

Carbon dye histologically confirms the identity of sentinel lymph nodes in cutaneous melanoma.

BACKGROUND: False-negative results from lymphatic mapping and sentinel lymphadenectomy (LM/SL) are associated with technical failures in nuclear medicine and surgery or with erroneous histologic evaluation. Any method that can confirm sentinel lymph node (SN) identity might decrease the false-negative rate. Carbon dye has been used as an adjunct to assist lymphadenectomy for some tumors, and the authors hypothesized that it could be used for the histologic verification of SNs removed during LM/SL. The current study assessed the clinical utility of carbon dye as a histopathologic adjunct for the identification of SNs in patients with melanoma and correlated the presence of carbon particles with the histopathologic status of the SNs. METHODS: LM/SL was performed using carbon dye (India ink) combined with isosulfan blue dye and sulfur colloid. Blue-stained and/or radioactive lymph nodes (two times background) were defined as SNs. Lymph nodes were evaluated for the presence of carbon particles and melanoma cells. If an SN lacked carbon dye in the initial histologic sections, four additional levels were obtained with S-100 protein and HMB-45 immunohistochemistry. Completion lymph node dissection (CLND) was performed if any SN contained melanoma cells. RESULTS: One hundred patients underwent successful LM/SL in 120 lymph node regions. Carbon particles were identified in 199 SNs from 111 lymph node regions of 96 patients. Sixteen patients had tumor-positive SNs, all of which contained carbon particles. The anatomic location of the carbon particles within these tumor-positive SNs was found to be correlated with the location of tumor cells in the SNs. The presence of carbon particles appeared to be correlated with blue-black staining (P = 0.0001) and with tumor foci (P = 0.028). All 35 non-SNs that were removed during LM/SL were tumor-negative, and only 2 contained carbon particles. Of the 272 non-SNs removed during CLND, 5 contained metastases; 3 of these 5 were the only non-SNs that had carbon particles. The use of carbon particles during LM/SL was found to be safe and nontoxic. CONCLUSIONS: Carbon dye used in LM/SL for melanoma permits the histologic confirmation of SNs. Carbon particles facilitate histologic evaluation by directing the pathologist to the SNs most likely to contain tumor. The location of carbon particles within SNs may assist the pathologist in the detection of metastases, thereby decreasing the histopathologic false-negative rate of LM/SL and subsequently reducing the same-basin recurrence rate.

Completely resected anaplastic thyroid carcinoma combined with adjuvant chemotherapy and irradiation is associated with prolonged survival.

BACKGROUND: The prognosis of anaplastic thyroid carcinoma (ATC) has been dismal. The objective of this study was to identify prognostic factors in patients who had prolonged survival. METHODS: Patients with ATC were identified from a computer database at a tertiary referral center. Univariate and multivariate analyses for survival differences were performed using the Kaplan-Meier log-rank statistic and the Cox proportional hazards model, respectively. RESULTS: Of the 33 evaluable patients, median survival was 3.8 months. Median age was 69 years. Prior goiter was present in 6 patients (18%), and 6 (18%) had prior thyroid carcinoma. Median tumor size was 6 cm, and 12 (36%) had adjacent well-differentiated carcinoma. Of the 26 patients who underwent neck exploration, 8 patients were potentially cured and received postoperative chemotherapy and irradiation; 4 (50%) were surgically macroscopically free of disease, and 4 (50%) patients had minimal residual disease after total thyroidectomy and resection of tumor adherent to adjacent structures. Four of these 8 patients survived longer than 2 years; their 5-year survival estimate was 50%. Eighteen patients underwent palliative resection of neck disease, leaving macroscopic residual disease or distant metastases; postoperative adjuvant chemotherapy and irradiation were administered in 16 of these 18 patients. Seven patients were treated with only chemotherapy and irradiation. In patients treated with potentially curative resection, median survival was 43 months compared with 3 months with palliative resection (P =0.002); the median survival of 3.3 months with only chemotherapy and irradiation was no different than palliative resection (P =0.63). No association was found between survival and age, prior goiter, prior thyroid carcinoma, adjacent differentiated carcinoma, or tumor size. CONCLUSIONS: Although the prognosis of most patients with ATC continues to be poor, complete resection of ATC combined with postoperative adjuvant chemotherapy and irradiation resulted in long-term survival, even with persistent minimal disease that remained on vital structures. An aggressive attempt at maximal tumor debulking followed by adjuvant therapy was found to be warranted in patients with localized ATC.

Application of an [(18)F]fluorodeoxyglucose-sensitive probe for the intraoperative detection of malignancy.

BACKGROUND: Whole-body positron emission tomography (PET) has been shown to be a highly sensitive method for detecting malignancy not imaged by conventional modalities. We have adapted a hand-held gamma-ray-sensitive probe to detect the radiation emission from the [(18)F]fluorodeoxyglucose (FDG) used in PET imaging. This pilot study was devised to examine the feasibility of using a hand-held probe to intraoperatively differentiate normal from tumor-bearing tissue. MATERIALS AND METHODS: A commercially available gamma probe was adapted to detect the radioactivity released from FDG and examined to determine the in vitro sensitivity for localization of a FDG point source. Eight consecutive patients underwent resection of metastatic colon cancer or melanoma; each received a preoperative injection of 7--10 mCi of FDG. The gamma probe was used to determine radioactive counts per second from tumor and normal tissue, and ratios of tumor to adjacent normal background were calculated. RESULTS: In vitro studies with a FDG point source demonstrated the probe could identify the source with a 50% reduction in maximum counts 1.7 +/- 0.1 cm from the source (full-width half-maximum measurement). Based on the results of their preoperative PET scans 17 tumors were identified from the 8 patients. Of the 17 tumors assessed the in vivo tumor-to-background ratios varied from 1.16:1 to 4.67:1 for the melanoma patients (13 tumors) and from 1.19:1 to 7.92:1 for colon cancer patients (4 tumors). Thirteen tumors were resected; four (2 patients) were unresectable. CONCLUSIONS: This study demonstrates the use of a hand-held gamma-ray-sensitive probe to intraoperatively differentiate the radioactivity released from FDG from tumor-bearing and adjacent normal tissue. While further studies are necessary for us to optimize the use of this probe, the intraoperative detection of FDG-avid malignancies may ultimately improve our ability to completely resect patients with metastatic disease.

Factors affecting sentinel node localization during preoperative breast lymphoscintigraphy.

UNLABELLED: Variable success rates for identifying axillary (AX) sentinel nodes in breast cancer patients using preoperative lymphoscintigraphy have been reported. We evaluated the effects of age, weight, breast size, method of biopsy, interval after biopsy, and imaging view on the success of sentinel node identification and on the kinetics of radiopharmaceutical migration. METHODS: Preoperative breast lymphoscintigraphy was performed in consecutive breast cancer patients from February 1998 to December 1998. The ipsilateral shoulder was elevated on a foam wedge and the arm was abducted and elevated overhead. Imaging using this modified oblique view of the axilla (MOVA) started immediately after peritumoral injection of Millipore-filtered 99mTc-sulfur colloid and continued until AX sentinel nodes were identified. Anterior views were obtained after MOVA. AX, internal mammary (IM), and clavicular (CL) basins were monitored in all patients. MOVA was compared with the anterior view for sentinel node identification. Age, weight, breast size, method of biopsy, interval after biopsy, and primary tumor location were evaluated for their effects on sentinel node localization and transit times from injection to arrival at the sentinel nodes. RESULTS: Seventy-six lymphoscintigrams were obtained for 75 patients. AX sentinel nodes were revealed in 75 (99%) cases. IM or CL sentinel nodes were found in 19 (25%) cases and were not related to tumor location; exclusive IM drainage was present in 1 (1%) case. Identification of AX sentinel nodes was equivalent with MOVA and anterior views in 18 (24%) patients, was better with MOVA in 20 (26%) patients, and was accomplished only with MOVA in 38 (50%) patients. Median transit time was 17.5 min (range, 1 min to 18 h) after injection, and larger breast size was associated with increased transit time. No effect of age, weight, biopsy method, interval from biopsy, or tumor location on transit time was found. CONCLUSION: Use of MOVA can improve identification of AX sentinel nodes. Although AX drainage is the predominant pattern, a tumor in any portion of the breast can drain to IM sentinel nodes. Transit time was influenced by breast size. Overall short arrival times with this technique allow sentinel lymph node dissection to be performed on the same day as lymphoscintigraphy.

Sentinel lymphadenectomy in node negative breast cancer.

SLND, regardless of method, can precisely predict the status of the axillary lymph nodes. Despite differences in technique, the consistent results support the sentinel node hypothesis in breast cancer. The procedure is well tolerated, and staging can be achieved accurately with minimal morbidity. SLND is a minimally-invasive procedure that provides tissue for the pathologist that represents the site most likely to harbor metastases. If a negative sentinel node is removed at SLND, it equates to truly node-negative breast cancer in almost all cases when done by experienced surgeons familiar with the technique. SLND can be mastered by surgeons at several institutions, but requires appropriate training to learn the technique. The team involved in SLND, which consists of the surgeon, pathologist and nuclear medicine physician, must determine its own false negative rate for the procedure, which requires a concomitant ALND so that accuracy is validated. Multicenter randomized clinical trials from the American College of Surgeons and NSABP are in progress, which will evaluate in general, although with different randomization schemes, the outcome of patients who have SLND alone compared to those who have ALND. Before ALND is completely abandoned, these trials must be completed so that the role of SLND in the management of all patients with early breast cancer is fully defined.

Origin of metallic particles resembling microcalcifications on mammograms after use of abrasive cautery-tip cleaning pads during breast surgery: experimental demonstration.

PURPOSE: To determine if the act of cleaning a cautery tip with an abrasive pad dislodges radiopaque particles that can be transferred to breast tissue during surgery, thereby mimicking microcalcifications at mammography. MATERIALS AND METHODS: Mock breast surgery was performed by cauterizing bovine liver or fresh, normal, human breast tissue. The cautery tip was rubbed against a cleaning pad five to 20 times in the manner used intraoperatively and was touched on separate breast tissue specimens two to six times. Specimen radiography was then performed. Thirty-six breast specimens were used in three experiments, including 28 used for the experimental conditions and eight control specimens. RESULTS: Particles collected from the cleaning pads resembled microcalcifications. After cauterization of liver, breast tissue, or both, in series, particles transferred from the cautery tip to breast tissue specimens could be identified on specimen radiographs. Transfer of particles after cautery of breast tissue occurred with increased numbers of rubs and specimen contacts. CONCLUSION: Radiopaque aluminum oxide particles from abrasive cautery-tip cleaning pads can be dislodged and transferred to breast tissue during surgery. Scrutiny of high-detail, spot-compression, magnification mammograms will help identify these particles. Simple measures to mitigate particle transfer during breast surgery can prevent this problem and obviate a potential second procedure to remove particles mistaken for microcalcifications.

Prospective observational study of sentinel lymphadenectomy without further axillary dissection in patients with sentinel node-negative breast cancer.

PURPOSE: Immediate complete axillary lymphadenectomy (ALND) after sentinel lymphadenectomy (SLND) has confirmed that tumor-negative sentinel nodes accurately predict tumor-free axillary nodes in breast cancer. Therefore, we hypothesized that SLND alone in patients with tumor-negative sentinel nodes would achieve axillary control, with minimal complications. PATIENTS AND METHODS: Between October 1995 and July 1997, 133 consecutive women who had primary invasive breast tumors clinically

Sentinel lymph node dissection for thyroid malignancy.

Sentinel lymph node dissection (SLND) for melanoma and breast cancer has been validated as an accurate technique to assess the status of the lymph nodes in the regional drainage basin. The sentinel node concept has also been investigated in other solid tumors, and more recently, in thyroid carcinoma. SLND using a vital blue dye during thyroidectomy for suspected thyroid malignancy successfully identifies sentinel nodes, with minimal morbidity. Excised sentinel nodes can be examined for micrometastases, and if negative, then the rest of the cervical nodes are likely to be negative. The false negative rate of SLND for thyroid malignancy is unknown, however, because modified neck dissections have not accompanied all cases. The impact that lymph node metastasis in thyroid carcinoma has on prognosis is debatable, unlike breast cancer and melanoma, which therefore makes the utility of thyroid SLND less clear. The technique, results, and morbidity of SLND during thyroidectomy is presented, and its possible utility in well-differentiated and medullary thyroid carcinoma is discussed.

Role of sentinel lymph node dissection in breast cancer.

The sentinel node concept is valid for penile cancer, melanoma, breast cancer and is probably also applicable to other solid malignancies. Sentinel nodes are the one or two initial nodes in the regional nodal drainage basin encountered by the lymphatic effluent from a tumour, which can be identified with an injection of vital dye or other lymphogogue. Sentinel lymph node dissection (SLND), a minimally invasive procedure with negligible morbidity, has therefore been utilized as an alternative to complete axillary lymph node dissection (ALND) for staging breast cancer. Examination of sentinel nodes provides a focused histopathological assessment of tissue most likely to harbour metastases, providing enhanced staging accuracy with a low false-negative rate. Tumour-free sentinel nodes are predictive of a tumour-free axilla, thereby allowing for the possibility of SLND without ALND and sparing patients the morbidity of ALND. Most of the experience from SLND has been obtained for axillary sentinel nodes. However, sentinel nodes have been identified in nonaxillary sites, such as the internal mammary nodes, but data on SLND for these regions is scarce. The ultimate role of SLND in breast cancer, which may be to identify sentinel-node-negative patients or even those with sentinel node metastases who can safely avoid ALND without sacrificing regional control and possibly gain a therapeutic benefit, cannot be defined before we have the results of large trials that are currently in progress.

Biopsy method and excision volume do not affect success rate of subsequent sentinel lymph node dissection in breast cancer.

INTRODUCTION: Sentinel lymph node dissection (SLND) is becoming a recognized technique for accurately staging patients with breast cancer. Its success in patients with large tumors or prior excisions has been questioned. The purpose of this study was to evaluate the effect of biopsy method, excision volume, interval from biopsy to SLND, tumor size, and tumor location on SLND success rate. METHODS: Consecutive patients who underwent SLND followed by completion axillary lymph node dissection from October 1991 to December 1995 were analyzed. Included were cases performed early in the series before the technique was adequately developed. Excision volume was derived from the product of three dimensions as measured by the pathologist. Two end points were analyzed: sentinel node identification rate and accuracy of SLND in predicting axillary status. Univariate analyses using chi2 or Fisher's exact test for categorical variables and Wilcoxon rank sums for continuous variables were performed. Multivariate analysis was performed using logistic regression. RESULTS: There were 284 SLND procedures performed on 283 patients. Median age was 55 years. The most recent biopsy method used before SLND was stereotactic core biopsy in 41 (14%), fine-needle aspiration in 62 (22%), and excision in 181 (64%) procedures. The mean excision volume was 32 ml with a range of 0.3-169 ml. The mean time from biopsy to SLND was 17 days with a range of 0-140 days. The mean tumor size was 2.0 cm (15 Tis [5%], 184 T1 [65%], 72 T2 [25%], and 13 T3 [5%]). Tumors were located in the outer quadrants in 74%, the inner quadrants in 18%, and subareolar region in 8%. The sentinel node was identified in 81%, and 39% had metastases. There were three false-negative cases early in the series. Sensitivity was 97%, and accuracy was 99%. Negative predictive value was 98% in cases in which the sentinel node was identified. On the basis of biopsy method, excisional volume, time from biopsy to SLND, tumor size, and tumor location, there was no statistically significant difference (P>.05) in sentinel node identification rate or accuracy of SLND. CONCLUSIONS: SLND has a high success rate in breast cancer patients regardless of the biopsy method or the excision volume removed before SLND. In addition, the interval from biopsy to SLND, tumor size, and tumor location have no effect on the success rate of SLND, even in this series which included patients operated on before the technique was adequately defined. Patients with breast cancers located in any quadrant and diagnosed either with a needle or excisional biopsy could be evaluated for trials of SLND.

Standardized probe-directed sentinel node dissection in melanoma.

BACKGROUND: Radiopharmaceutical agents appear to improve the accuracy of sentinel node (SN) identification in patients with early-stage melanoma, but the optimal radiopharmaceutical agent and its timing from injection to surgery remain controversial. We undertook this investigation to examine the utility of 3 methods of intraoperative lymphatic mapping with radiopharmaceutical-directed sentinel lymphadenectomy (LM/SL). We suspected that concurrent injection of radiopharmaceutical and blue-dye would lead to the greatest success of SN identification. METHODS: The study was composed of 247 consecutive patients who had American Joint Committee on Cancer stage I or II melanoma. Before LM/SL, all patients underwent cutaneous lymphoscintigraphy by 1 of 3 techniques: technetium 99m (Tc 99m) human serum albumin (HSA) injected at least 24 hours before LM/SL (124 patients), Tc 99m filtered sulfur colloid (SC) injected no more than 4 hours before LM/SL (same-day SC, 95 patients), or Tc 99m SC injected at least 18 hours before LM/SL (prior-day SC, 28 patients). At the time of LM/SL, isosulfan blue dye was injected alone (SC groups) or with a second dose of HSA (HSA group). A hand-held gamma probe was used to determine the radioactive (hot) counts of blue-stained and nonstained nodes, and the in vivo and ex vivo node-to-background count ratios of the nodes were compared. RESULTS: Preoperative LS identified 299 drainage basins; LM/SL identified at least 1 SN in 119 (98%) of 121 basins using same-day SC, 142 (97%) of 146 basins using HSA, and 32 (100%) of 32 basins using prior-day SC. There was no difference (P = .62) in the accuracy rate between the 3 techniques. The total number of SNs was 463. Same-day SC yielded higher intraoperative node-to-background count ratios than did either of the other techniques (P < .0001). Same-day SC also had the greatest relative change in radioactivity between the blue sentinel node and the post-excision basin (P < .0001), and the highest rate of SNs that were both blue and hot (in vivo or ex vivo ratio > or = 2, P = .05). CONCLUSIONS: LS and LM/SL performed on the same day with a single injection of filtered Tc 99m SC serves as the most useful method for probe-directed LM/SL. This technique demonstrated the highest in vivo and ex vivo count ratios, fall-off of radioactivity between the excised nodes and post-excision basin, and concordance between blue dye and hot nodes. It should be recommended as the method of choice for probe-directed LM/SL.

Accuracy of gamma probes in localizing radioactivity: in-vitro assessment and clinical implications.

BACKGROUND: Radioguided surgery (RGS) uses gamma probes to localize radioactive targets. We examined the in-vitro localizing properties of commercially available gamma probes for three common RGS radioisotopes: technetium-99m (99mTc), indium-111 (111In), and iodine-125 (125I). METHODS: The linear and angular localizing properties of five gamma probes were assessed. Radioactive counts in air were obtained at multiple energy threshold settings for each radioisotope. The full width half maximum (FWHM) and angular half maximum (AHM) were calculated to determine localizing accuracy. RESULTS: The FWHM of the five probes ranged from 1.2 to 3.4 cm for 99mTc, 1.8 to 2.7 cm for 111In and 1.6 to 2.8 cm for 125I. AHM of the probes ranged from 20 to 49 degrees for 99mTc, 14 to 35 degrees for 111In, and 18 to 47 degrees for 125I. The ranking of probe accuracy depended on the radioisotope studied. Absolute count rates varied with energy threshold settings for each probe and radioisotope. Probes required individualized energy threshold settings for optimal performance. CONCLUSIONS: The five probes varied in their characteristics for localizing the three radioisotopes. Each probe has particular detection characteristics that may change with different radioisotopes. Differences in radioisotope detection may explain intraoperative variation in localization during lymphatic mapping and sentinel lymph node dissection (LM/SL). Surgeons should be aware of probe performance characteristics before intraoperative use.

Anaplastic thyroid carcinoma.

Anaplastic thyroid carcinoma (ATC) is a rare, highly aggressive tumor characterized by rapid growth that causes death soon after diagnosis. Optimal treatment is debatable because of inherent difficulties investigating a rare malignancy associated with a survival of only months. Nevertheless, there is evidence that a multimodality approach consisting of surgery, chemotherapy, and radiotherapy is the preferred approach when the disease is localized. Eradication by complete surgical resection with total thyroidectomy, although often not feasible, followed or preceded by concurrent doxorubicin-based chemotherapy and hyperfractionated external beam radiation has been occasionally used and can rarely be associated with long-term survival. Surgical extirpation with radical organ resection is associated with high morbidity, and, although an aggressive surgical approach is reasonable in selected patients, organ preservation should be the goal in a disease with high recurrence and death rates. Any modality used in isolation usually fails to control local and regional ATC. In otherwise healthy patients who can tolerate aggressive therapy, chemoradiation and surgery may also be warranted for palliating incurable disease in the neck, which can cause death by asphyxiation.

Vaccine therapy for patients with melanoma.

Investigation into the therapeutic use of vaccines in patients with metastatic melanoma is critically important because of the lack of effective conventional modalities. The most extensively studied melanoma vaccines in clinical trials are whole-cell preparations or cell lysates that contain multiple antigens capable of stimulating an immune response. Unfortunately, in the majority of studies, immune responses to these vaccines have not translated into a survival advantage. Advances in tumor cell immunology have led to the identification of candidate tumor cell antigens that can stimulate an immune response; this, in turn, has allowed for refinements in vaccine design. However, the exact tumor antigens that should be targeted with a specific vaccine are unknown. The univalent antigen vaccines, which have greater purity, ease of manufacturing, and reproducibility compared with polyvalent vaccines, may suffer from poorer efficacy due to immunoselection and appearance of antigen-negative clones within the tumor. Novel approaches to vaccine design using gene transfection with cytokines and dendritic cells are all promising. However, the induction of immune responses does not necessarily confer a therapeutic benefit. Therefore, these elegant newer strategies need to be studied in carefully designed clinical trials so that outcomes can be compared objectively with standard therapy. If survival is improved with these vaccine approaches, their ease of administration and lack of toxicity will firmly entrench active specific vaccine immunotherapy as a standard modality in the treatment of the melanoma patient.

Long-term survival after complete resection of melanoma metastatic to the adrenal gland.

BACKGROUND: Survival of patients with American Joint Committee on Cancer stage IV melanoma is generally poor, although there are occasional long-term survivors who have undergone surgical resection of a limited number of metastases. In the study, we examined the outcome of patients with adrenal gland metastases. METHODS: Eighty-three patients with adrenal metastases were identified from our computerized melanoma database of 8250 patients. Univariate and multivariate analyses for overall survival differences were performed by using proportional hazards modeling. RESULTS: Median survival for the 83 patients was 9.3 months (1-67 months). Of the 27 patients who underwent surgical exploration, 18 (66%) were rendered clinically free of disease by adrenalectomy alone (12 cases) or by adrenalectomy and resection of additional disease (6 cases). Nine patients underwent palliative adrenal resection. Median survival was 25.7 months after complete resection compared with 9.2 months after palliative resection (P = .02). CONCLUSIONS: Patients with adrenal metastases from melanoma, either isolated or with a limited number of additional metastases, may benefit from surgical resection if all visible disease can be removed. Patients with unresectable extra-adrenal disease achieve no survival benefit from adrenalectomy.