Unknown Primaries Research Articles

Two Cases of Bilateral Tonsillar Carcinoma with Extensive Metastasis

Received January 11, 2013 Revised March 19, 2013 Accepted April 5, 2013 Address for correspondence Jae-Gu Cho, MD, PhD Department of OtorhinolaryngologyHead and Neck Surgery, Korea University College of Medicine, 148 Gurodong-ro, Guro-gu, Seoul 152-703, Korea Tel +82-2-2626-3189 Fax +82-2-868-0475 E-mail jgcho@korea.ac.kr Tonsillar carcinoma is the most common cancer of oropharynx. Panendoscope guided biopsy has been estabilished as a diagnostic and therapeutic procedure for patients with cervical nodes from metastasis of unknown origin. There is a controversy about regarding bilateral tonsillectomy as the standard procedure. In studies where bilateral tonsillectomy are routinely performed for diagnosis of unknown primaries, the rates of occult tonsillar carcinoma could be as high as 40%. On the other hand, the incidence of synchronous tonsil carcinoma is exceedingly rare with less than 15 reports in the English literature although synchronous, metachronous carcinomas are more common in head and neck carcinomas than in other sites because the risk factors for the development of carcinomas are shared between the two. We report two cases of bilateral tonsillar carcinoma with literature reviews. In these particular cases, they present an aggressive clinical feature despite small size of the primary lesion. A careful, systemic work up should be performed in bilateral tonsillar carcinoma. Korean J Otorhinolaryngol-Head Neck Surg 2013;56:651-4 Key WordsZZBilateral ᆞCarcinoma ᆞPalatine tonsil ᆞSynchronous.

A new paradigm for the diagnosis and management of unknown primary tumors of the head and neck: A role for transoral robotic surgery

In 2% to 3% of patients with cancer metastatic to cervical lymph nodes, a primary tumor will not be found despite exhaustive diagnostic efforts. The treatment for these patients includes cervical lymphadenectomy followed by radiation to areas with increased risk of harboring a mucosal primary. Wide-field radiation therapy increases the incidence of xerostomia and dysphagia. Localizing a primary tumor has thus both therapeutic and quality-of-life implications, allowing possible complete surgical excision, concentrated radiation therapy, and potential deintensification of adjuvant therapy. With improved visualization and freedom of motion, transoral robotic surgery (TORS) is an innovative surgical modality that allows resection of oropharyngeal subsites with minimal morbidity. Retrospective chart review. Ten patients with unknown primary tumors of the head and neck were identified. All patients underwent a cervical biopsy, positron-emission tomography/computed tomography, formal endoscopy, and bilateral tonsillectomy. When the initial endoscopy and biopsies did not localize a primary tumor, all patients underwent transoral robotic base of tongue resection. Evaluation of the patients' oropharyngeal mucosa using the robot did not reveal an obvious lesion and no palpable tumors were appreciated in the resected specimens. In 9/10 (90%) patients, pathologic examination revealed invasive squamous cell carcinoma (SCCA) with a mean diameter of 0.9 cm. Unknown primary SCCA presents a diagnostic challenge to the head and neck surgeon. We present a small series of tumors that would have been treated as unknown primaries under traditional diagnostic and therapeutic algorithms. TORS base of tongue resection identified primary tumors in 90% patients with minimal morbidity.

P16 Expression in cutaneous squamous carcinomas with neck metastases: A potential pitfall in identifying unknown primaries of the head and neck

Human papillomavirus (HPV) positivity (+) has been used to identify oropharyngeal squamous carcinomas (SCCs) presenting as unknown primaries in the neck. p16 overexpression correlates with HPV+ in the oropharynx; however, the use of p16 alone as a surrogate marker of oropharyngeal HPV+ tumors has not been validated. We immunohistochemically analyzed p16 expression in surgically resected aggressive cutaneous head and neck SCC primaries and their nodal metastases from 24 patients to determine the potential overlap of p16 expression outside of the oropharynx. Five of 24 primary tumors (20.8%) and 3 lymph node metastases (12.5%) in levels II, III, and V, and the periparotid region diffusely expressed p16. HPV (high-risk types by in situ hybridization) was negative. p16 expression is relatively common in lymph node-positive cutaneous head and neck SCCs; thus, p16 expression as an independent biomarker and mechanism to determine the oropharyngeal source of an unknown primary is not advised.

A SEER-based Probability Model Using HPV Status and Lymph Node Levels to Predict Primary Head-and-Neck Tumor Site for Squamous Cell Cancer Confined to Cervical Lymph Nodes

To develop a multifactorial statistical model to predict the specific head and neck (H&N) tumor site origin in cases of squamous cell carcinoma (SCC) confined to the cervical lymph nodes (“unknown primaries”) using recent biomarker and epidemiologic data.

Cancer of unknown primary (CUP): does cause of death and family history implicate hidden phenotypically changed primaries?

Cancer of unknown primary (CUP) is diagnosed at the metastatic stage. We aimed to identify hidden primary cancers in CUP patients by comparison with cancers in family members. We take use of the fact that the cause of death in CUP patients is often coded as the cancer in the organ of fatal metastasis. Forty-one thousand five hundred and twenty-three CUP patients were identified in the Swedish Family-Cancer Database, and relative risks (RRs) were calculated for cancer in offspring when family members were diagnosed with CUP and died of the cancer diagnosed in offspring. The RR for lung cancer in offspring was 1.85 when a family member was diagnosed with CUP and died of lung cancer. Significant familial associations were found for seven other cancers. Many familial associations were also significant when offspring CUP patients died of the cancer diagnosed in family members. The cause of death after CUP diagnosis frequently matched the cancer found in a family member, suggesting that the CUP had originated in that tissue. The metastasis had probably undergone a phenotypic change, complicating pathological tissue assignment. These novel data suggest that some CUP cases are phenotypically modified primary cancers rather than cancers of unknown primaries.

The Who and When of Molecular Testing for Tumors of Unknown Primaries: One Resident’s Perspective

To the Editor It was with great interest that I read the article by Dumur and colleagues1 on their clinical verification of the only US Food and Drug Administration–approved microarray-based test for the diagnosis of poorly and undifferentiated tumors, the Pathwork Tissue of Origin (TOO) Test (Pathwork Diagnostics, Sunnyvale, CA). Briefly, the Pathwork TOO Test is one of several tests currently on the market (CancerTYPE ID, Biotheranostics, San Diego, CA, and miRview mets, Rosetta Genomics, Rehovot, Israel, being others) that attempt to correlate the RNA profile of a given undifferentiated tumor with its suspected tissue of origin (hence, the clever names) through complementary DNA microarray analysis and predetermined algorithms. Based on the findings in the study by Dumur et al,1 the Pathwork TOO test seems to perform glowingly, with a reported 97% concordance rate between the TOO result and final diagnosis, when immunohistochemical and radiologic studies are included. Yet, while most of the available data for these assays are and have been promising with regard to their ability to retrospectively correlate diagnoses with standard analytic methods, several important questions remain, including “When is it appropriate to use one of these tests?” and “Who decides when these tests should be ordered?” As these tests are relatively new to the market and the field of pathology in general, some basic questions are, of course, warranted, especially because they are rather costly, and, as yet, there have been no published prospective studies comparing ultimate outcomes of cases diagnosed with these new techniques vs cases undergoing standard immunohistochemical workups. A recent case seen in our group raises some of these questions and provides a real-life example of when requests for these assays could be made. The patient was a 69-year-old man who had an abdominal computed tomography (CT) scan following a 1-month …

The role of PET-CT in detecting unknown primary tumour in patients with cervical lymph node metastases

PET-CT examination was conducted with 440 patients treated at the Department of Head and Neck Surgery, National Institute of Oncology, Budapest, between January 1, 2006 and December 31, 2010. Out of them 77 patients were selected with whom no examination of any sort (physical, pan-endoscopy, or any of the conventional imaging techniques) succeeded in identifying the primary tumour. In each case the primary examination (aspiration cytology and histology) verified cervical metastases, most of them being squamous cell carcinoma. The significance of PET-CT was retrospectively evaluated in cases of unknown primary tumour with verified cervical metastases. We tested the sensitivity of PET-CT in detection of the primary malignant tumour, and possible distant metastases or a second primary in order to plan an optimal treatment schedule for the patient. Patients with whom the examinations specified in the treatment protocol (physical examination, pan-endoscopy, conventional imaging, biopsy) had failed to diagnose the primary tumour were referred to PET-CT. In each case 18F-FDG tracer was used. In 21/77 patients (27%), the PET-CT yielded unequivocal evidence for the primary tumour confirmed by histology, as well. With 10 others (13%), the precarious diagnoses by various imaging techniques were confirmed by the PET-CT. False positive findings with PET-CT that were not verified either by histology or control examination tests occurred but in 10 patients (13%). Concerning the primary tumour, false negative result was obtained only with 3 patients (4%). It should be noted that their retrospective evaluation proved diagnostic errors, the primary tumours were visible in all the scans. With 33 patients (43%) PET-CT furnished no additional information compared to the previous examinations. In 10 patients, asymptomatic distant metastases and in 3 patients synchronous tumours were diagnosed. We also acknowledge that the significance of PET-CT using 18F-FDG is unquestionable in the detection of unknown primary tumours. It is strongly recommended to re-include the detection of unknown primaries in the approved national indication list of PET-CT. (Note, until January 1, 2008 it had been included!) PET-CT is capable of detecting a primary tumour, after all unsuccessful diagnostic examinations till then, in 25-40% of the cases. One cannot disregard the role and significance of PET-CT in the detection of asymptomatic synchronous tumours, or distant metastases. These benefits make PET-CT a suitable tool for the refinement of individually tailored treatment strategies leading to better therapeutic results and more favourable cost-benefit ratio.

Value of fused positron emission tomography CT in detecting primaries in patients with primary unknown cervical lymph node metastasis

Identification of the primary tumour can prolong the life expectancy of patients with primary unknown cervical lymph node metastasis (PUCLNM) through targeted therapy. This study investigated the value of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET-CT) at identifying primaries in patients with PUCLNM. Twenty-seven patients (21 males and 6 females, median age 48.2 ± 16.3, age range 30-73) with PUCLNM underwent FDG PET-CT to search for the primary tumour, which could not be detected by conventional diagnostic modalities. The results were analysed and correlated with either pathological findings or clinical follow up. Pathological FDG uptake suspicious for the primary was detected in 13 cases, while the primary tumour remained occult in 14 cases. Eleven of 13 patients with suspected primaries were confirmed by histological findings. One with a coexisting second tumour and three with unexpected distant metastases were found in patients with confirmed primaries. The most common primary location in patients with PUCLNM found in our study was nasopharynx. In those 14 patients with negative FDG PET-CT results, only one patient had a primary malignancy that was proven histologically after endoscopy with biopsy during a period of clinical follow up. The sensitivity, specificity, accuracy and positive predictive values of FDG PET-CT were 91.7, 86.7, 88.9 and 84.6%, respectively. FDG PET-CT is a useful tool to help search for unknown primaries in patients with cervical lymph node metastasis and has an acceptable diagnostic yield for the detection of distant malignancies.

Noncutaneous Melanoma Have Distinct Features from Each Other and Cutaneous Melanoma

Objective: Compared to the cutaneous melanoma, noncutaneous melanomas are relatively rare and have a distinct pattern of behavior. We performed this study to define the clinical characteristics and outcomes of patients with noncutaneous melanomas and emphasize how they differ from cutaneous melanomas. Methods: 216 patients who were diagnosed with noncutaneous melanoma were assessed and their medical records between 2000 and 2010 were retrieved from the cancer registry. 475 patients with cutaneous melanoma were used for comparison. Results: Of the 216 noncutaneous melanoma patients, 83 had mucosal melanoma, 101 had ocular melanoma and 32 had unknown primaries. For mucosal melanoma, the head and neck was the most common anatomic site (53%), followed by the gastrointestinal tract (37%), female genital (6%) and urinary tract (4%). The majority of the ocular melanomas (94%) originated in the uvea. 32 patients demonstrated a primary unknown disease with regional metastasis as the dominant behavior (72%). Age was found to be statistically significantly different among melanoma patients (p < 0.001). Younger patients had more cutaneous and unknown primary melanomas, whereas older patients had more ocular and mucosal melanoma. In subset analyses, we found significant differences between cutaneous and ocular (p = 0.038) or mucosal (p < 0.001) melanomas. The ratios of metastasis on admission were higher in patients with mucosal (27.7%) and unknown primary (28.1%) melanomas, and lower for ocular (3%) melanomas (p < 0.001). Patients with cutaneous melanoma had an intermediate (12%) ratio. Patients with ocular melanoma had better outcome than patients with other melanoma sites (p < 0.05). While overall survival for cutaneous melanoma was significantly negatively correlated with male gender (p < 0.001), advanced stages (p < 0.001) and old age (p = 0.005), stage IV disease was the only independent prognostic factor in patients with ocular melanoma (p = 0.006). No prognostic factor for outcome was found for either mucosal or unknown primary melanomas. Conclusion: The clinical presentations and prognoses of different primary sites of melanoma are distinctive. Therefore, approaches to these melanomas should be different.

Transoral robotic biopsy of the tongue base: A novel paradigm in the evaluation of unknown primary tumors of the head and neck

Squamous cell carcinoma of the head and neck can present as a cervical metastasis from an unknown primary site. The standard diagnostic workup includes panendoscopy and directed biopsies but this will fail to identify a portion of unknown primary tumors. Herein, we present a case report of a male patient with an unknown primary tumor in which the da Vinci surgical robot was used to evaluate the tongue base. Clinical evaluation, imaging, and panendoscopy with directed biopsies failed to detect the primary tumor site. Robot-assisted biopsy of a broad area of the tongue base, incorporating submucosal tissue, identified the primary tumor with minimal postoperative morbidity. Failure to localize an unknown primary tumor often results in widespread irradiation of the upper aerodigestive tract, inducing significant morbidity. Robot-assisted biopsies of the tongue base may identify unknown primaries that would otherwise have been missed through standard directed biopsy techniques.

Aflibercept (VEGF Trap) in Inoperable Stage III or Stage IV Melanoma of Cutaneous or Uveal Origin

Aflibercept is a soluble decoy VEGF receptor and angiogenesis inhibitor with potent preclinical antitumor activity in melanoma. We conducted a multicenter phase II study in patients with inoperable stage III or IV melanoma and no prior chemotherapy. A two-stage design was adopted to evaluate 4-month progression-free survival rate (PFSR) and response rate. Aflibercept was given at 4 mg/kg intravenously every 2 weeks. Response was assessed every 8 weeks. First-stage accrual of 21 patients was specified and with an adequate 4-month PFSR accrual continued to a total of 41. Forty-one patients of ages 23 to 84 (median = 57) were enrolled. Thirty-nine had American Joint Committee on Cancer stage IV (5 M1a, 7 M1b, and 27 M1c) and 2 had inoperable stage IIIC (N3). Eastern Cooperative Oncology Group (ECOG) performance status was 0 (27 patients) or 1 (14 patients). Ten patients had primary uveal melanoma, 28 cutaneous, and 3 had unknown primaries. A median of 7 cycles were initiated (range: 1-56). Grade 3 and 4 toxicities included hypertension in 9 patients (22%) and proteinuria in 6 (15%). Among 40 patients evaluable for efficacy (those who initiated aflibercept), 3 (7.5%) had a confirmed partial response and 20 had progression-free survival of 4 months or above. The predicted 1-year survival rate derived from the Korn meta-analysis model is 36% (N = 39), whereas we observed a corresponding 56.4% survival rate at 1 year (95% CI, 43-74, P < 0.005). Median overall survival in this trial is 16.3 months (95% CI, 9.2 to not reached). We observed a significant association between severity of hypertension following aflibercept and survival improvement. Aflibercept showed promising activity in patients with metastatic melanoma of cutaneous or uveal origin. Further evaluation of aflibercept as a single agent and in combination is warranted.

Role of Positron Emission Tomography-Computed Tomography in Head and Neck Cancer

Head and neck squamous cell carcinoma (HNC) is the sixth most common cancer worldwide. In the United States, it accounts for approximately 2% of all cancers and 2% of cancer deaths. The introduction of integrated positron emission tomography and computed tomography (PET/CT) has revolutionized imaging by permitting improved and more accurate anatomic localization of functional abnormalities in the complex territory of the head and neck region, and PET/CT has become a standard clinical imaging modality in patients with HNC. The main indications for PET/CT in HNC are in pretherapy staging, detection of unknown primaries, and monitoring of therapy response or disease surveillance. Although PET/CT is a promising tool in diagnosis and surveillance of HNC, there is lack of consensus as to its use, accuracy, and implications for patient management. The existing literature on the role of PET/CT in the management of HNC is reviewed, and a summary of the current debate is provided. Second primary cancers are the main cause of death among HNC patients with early disease, and the presence of distant metastases greatly impairs the survival of patients with advanced HNC. Therefore, early detection of second primary and metastatic tumors is imperative for optimizing survival outcome. However, given the lack of randomized, prospective trials addressing the role of PET/CT after chemoradiotherapy, the ideal function of PET/CT in disease surveillance has yet to be defined.

10th International Symposium on head and neck skin cancer

Since 1993, ten multidisciplinary symposia were organized at The Netherlands Cancer Institute on the diagnosis and treatment of malignancies of the head and neck. The symposia are meant to provide up-to-date teaching for physicians by world-renowned speakers. The previous symposia dealt with sarcomas, reconstruction, cancer in young patients, salivary glands, melanoma, unknown primaries, as well as several other topics. This 10th symposium focused on skin cancer of the head and neck. There are many types of skin cancer and the differential diagnosis can often be difficult. In this symposium, diagnosis, molecular biology, epidemiology, staging and the treatment of various skin cancers were discussed by leaders in the field. There were over 200 participants from many different countries in Europe and overseas, representing specialties in the fields of dermatology, maxillofacial surgery, otolaryngology, head and neck surgery, general surgery, plastic and reconstructive surgery, and radiotherapy.

Transoral laser microsurgery: A new approach for unknown primaries of the head and neck

To evaluate the efficacy of transoral laser microsurgery (TLM) used at examination under anesthesia (EUA) for detection and management of an unidentified primary site and to determine survival with both TLM EUA and traditional rigid pharyngolaryngoscopy EUA, with directed biopsies. Comparative retrospective review of patients who underwent two different procedures, TLM EUA and traditional EUA, to identify a primary site. Thirty patients presenting with occult primary met the study criteria. Eighteen underwent TLM EUA and 12 underwent traditional EUA. We collected data on the treatment approach, detection rate of the primary site, neck dissection, postoperative radiotherapy, and disease-free survival (DFS). The primary site detection rate with TLM EUA was 94% (17 of 18) and with traditional EUA was 25% (3 of 12). Overall, the occult primary was identified in 20 of 30 patients. The majority of patients (95%) had a primary in the oropharynx (19 of 20). Sixteen occult primaries in the TLM EUA group were immediately resected with TLM. At median follow-up of 30 months, there was no recurrence in the TLM EUA group. There was a 41.6% (5 of 12) recurrence rate in traditional EUA group. The Kaplan-Meier 5-year DFS was 100% for the TLM EUA group and 44% for the traditional EUA group (log rank value = 0.0006). TLM management of occult primary malignancies allowed high detection rates of primary tumor and was associated with a high level of DFS. Application of TLM during EUA both detects and treats the primary and may decrease the number of patients requiring wide-field irradiation.

Metastatic vs primary malignant neoplasms affecting the umbilicus: clinicopathologic features of 77 tumors

Periumbilical skin is unique due to its proximity to intra-abdominal and pelvic structures. In addition to primary skin malignancies, it is a site often involved with metastatic disease. We reviewed the clinical and pathologic features of 77 umbilical malignancies occurring at our institution since 1988. Seventy-seven patients were identified (female/male ratio, 4.1:1.0) with the average age for women being 63 years and 55 years for men. Eighty-eight percent of malignancies originated outside the umbilicus and 12% were primary skin tumors. Fifty-eight (85%) patients with metastatic tumors had umbilical involvement from a known primary vs 10 (15%) with unknown primaries. Nine patients with metastatic tumors to the umbilicus would present with solitary umbilical involvement. Of these patients, 56% would not have a primary site assigned to their metastatic disease. In women, the 3 most common primary sites were the ovary, endometrium, and pancreatobiliary tree, whereas for men, it was the genitourinary tract, pancreatobiliary tree, and the gastrointestinal tract. Of the primary umbilical malignancies, 44% of patients were male and 56% female. Malignant melanoma was the most common primary umbilical malignancy. In summary, women are more likely than men to have malignant tumors affecting the umbilicus. Overall, the most likely primary site of a metastatic tumor to the umbilicus is the genitourinary tract. Rarely, patients present with metastatic tumors to the umbilicus, and most of these patients will not have a primary site of tumor origin assigned.

Extracranial metastatic patterns on occurrence of brain metastases

Extracranial metastases and their frequency by sites have been described as prognostic factors for survival of patients with brain metastasis. However, these factors must be identified and described in more detail for a large series of patients. Using routine data from the largest German health insurance fund, 5,074 patients with brain metastasis who were diagnosed and treated in 2008 were analyzed to identify the frequency and distribution of extracranial metastatic sites concurrent with brain metastasis in relation to age, gender, and tumor type. Brain metastases were observed in males more frequently than in females (56.4 and 43.6% respectively P<0.001), and were most often from lung (51.2%), breast (12.3%), and unknown (7.5%) primaries. Extracranial metastatic sites were observed in 58.8% of patients; the number of sites was from 1 to 7, with a mean of 1.11. For the 16 most common primary sites the range was from 0.13 to 1.91 . In 11 of these 16 sites, lungs were the most common concurrent metastatic site. Lung cancer, breast cancer, non-Hodgkin's lymphoma, and testicular cancer most commonly metastasized to bone, and bladder cancer to kidneys. Different primary tumors have different frequencies and patterns of extracranial metastatic sites concurrently with brain metastasis. The lung is the most common metastatic site of most primary tumors, bone for a few tumors, and kidneys for bladder cancer. For the unknown primary tumor type, screening for these most common metastatic sites must be intensified, in particular when molecular assessment is not available.

Biopsy of Vertebral Tumour Metastasis for Diagnosing Unknown Primaries

To examine patients with vertebral tumour metastasis using transpedicular biopsy for diagnosing unknown primary tumours. 13 men and 8 women aged 41 to 80 (mean, 61) years with vertebral tumour metastasis of unknown primary origin underwent transpedicular biopsy of the affected vertebra. The origins of the primary tumours were lung cancer (n = 6), prostate cancer (n = 5), colorectal cancer (n = 5), kidney cancer (n = 4) and lymphoma (n = 1). All the specimens matched pathological characteristics of their corresponding primary tumours, except in one patient. This 42-year-old man had stage-4 colon cancer, in whom the pathologic findings could not enable differentiation between colon and prostate cancer. Transpedicular biopsy of the vertebra is a cost-effective diagnostic tool for evaluating unknown primary tumours.

New techniques to detect unknown primaries in cervical lymph node metastasis

Various methods have been reported for the detection of unknown primaries in cervical lymph node metastasis. Recently, we applied new optical devices and modifications of endoscopic techniques for the detection of primary lesions, and obtained excellent results. The detection rate of the new method was compared with that of previous methods. Retrospective. A total of 51 patients with cervical lymph node metastasis from an unknown primary site (CUP) were referred to our hospital between January 2000 and May 2009, and were retrospectively analyzed. Between 2000 and 2005, the observation by normal video-endoscope in straight head position was performed to detect the primary lesions in the outpatient setting. Since 2006, a new method for detection of primary lesions has been employed. The method includes the use of new optical devices (hooded video-endoscope and narrow-band imaging endoscope) and different head positions (head torsion technique, Valsalva maneuver, and the Killian position). The detection rate of primary lesions using the new method was 71% (15 of 21), which was better than the 30% (10 of 30) obtained with the conventional method. All primary lesions identified using the new method were located in the hypopharynx. The new method was able to detect primary lesions in 71% of cases with CUP. The higher rate achieved with the new method was probably due to the clear visualization of the hypopharynx. Thus, the new method was shown to be useful for the detection of primary lesions in cases of CUP.

Should unknown primary melanomas be excluded from adjuvant therapy trials? Insights from SWOG S0008.

8517 Background: Patients (pts) with unknown primary melanoma in regional nodes are widely considered to have a better prognosis, even adjusting for other factors, so many adjuvant trials exclude these pts. Methods: 432 stage IIIB/C pts with known primary (KP) cutaneous and unknown primary (UNK) melanoma went on S0008, a randomized trial of interferon-α2b vs. biochemotherapy. Pooled data blinded to treatment were analyzed for demographics and prognosis of KPC vs UNK. Pts with KP were divided as clinically evident (macromets) versus diagnosed by sentinel node biopsy, and by time from primary diagnosis (dx) to development of nodal metastases (0-6 mos; 6 mos-5 yrs; > 5 yrs). Since UNK pts by definition had macromets, comparative analyses involved only KPC pts with macromets. Results: Of 432 pts entered onto S0008, 33 (7.6%) were UNK, while 222 had KCP macromets. Thus, 13.1% of pts with macromets were UNK. Median age (48) and number of involved nodes (2) was the same for KCP macromets and UNK pts. Overall survival (OS) was not significantly different for the two groups (p = 0.46), but relapse-free survival (RFS) was significantly better for UNK (p = 0.05). When KCP pts were subdivided by time from primary to dx of macromets, RFS was significantly worse for pts whose metastases developed 0-6 mos from dx (p = 0.008), with a trend for worse OS (p = 0.08). RFS and OS for UNK pts were identical to KPC pts where the nodal metastases developed > 6 mos after primary dx. Conclusions: In S0008, unknown primary melanoma pts had significantly better RFS but not OS than pts with macromets from known primaries. However, their outcome was similar to known primaries diagnosed > 6 mos prior to presentation with nodal metastasis. Our data strongly support the inclusion of these pts in adjuvant therapy trials and do not suggest an intrinsically better (e.g., immunologically-mediated) outcome for pts with unknown primaries. Author Disclosure Employment or Leadership Position Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration Novartis, Schering-Plough Novartis, Schering-Plough Schering-Plough

Combination of MUC5ac and WT‐1 immunohistochemistry is useful in distinguishing pancreatic ductal carcinoma from ovarian serous carcinoma in effusion cytology

Malignant ascites may be the first presentation of an unsuspected cancer. Pancreas and ovary are among the organs that are usually evaluated as a source of primary. The purpose of this study is to investigate a panel of immunohistochemical stains to help differentiate pancreatic from ovarian carcinoma. We evaluated the immunohistochemical staining of eight commercially available antibodies MUC1, MUC2, MUC5ac, Wilm's tumor susceptibility gene 1 (WT1), cytokeratin 7 (CK7), CK20, CA125, and CA19.9 in 25 effusion specimens with evidence of metastatic carcinoma including 14 ovarian serous carcinomas, 9 pancreatic adenocarcinomas, and 2 unknown primaries. Primary ovarian serous carcinomas were positive for WT-1 (100%), CK7 (93%), CK20 (43%), CA125 (100%), CA19.9 (50%), MUC1 (100%), MUC2 (0%), and MUC5ac (0%). Primary pancreatic carcinomas were positive for MUC5ac (100%), MUC1 (100%), CA19.9 (100%), CK7 (78%), CK20 (22%), CA125 (89%), WT-1 (0%), and MUC 2 (0%). The combination of MUC5ac positivity/WT-1 negativity was seen in 100% of pancreatic carcinoma, whereas MUC5ac negativity/WT-1 positivity in 100% of ovarian serous carcinoma. It appears that the combination of MUC5ac and WT-1 stains is useful in distinguishing pancreatic ductal from ovarian serous carcinoma in body fluid cytology.