Cutaneous squamous cell carcinoma in patients with chronic lymphocytic leukemia

Patients with chronic lymphocytic leukemia are at heightened risk for the development of cutaneous malignancies, especially squamous cell carcinoma. These tumors portend a worse prognosis than is typically seen in immunocompetent patients. Patients with chronic lymphocytic leukemia who develop squamous cell carcinoma are more likely to develop multiple tumors, recurrences after treatment, and regional or distant spread of their tumor. Evaluation of the literature substantiates the ominous nature of squamous cell carcinoma in patients with underlying lymphocytic malignancies.     

Lymphocytic Infiltrates and Subclinical Epithelial Tumor Extension in Patients With Chronic Leukemia and Solid-Organ Transplantation

BACKGROUND: Dense infiltrates in association with squamous cell carcinoma (SCC) or basal cell carcinoma (BCC) in patients with underlying chronic lymphocytic leukemia (CLL) may complicate pathologic interpretation of histologic margins. OBJECTIVE: The study was conducted to determine the frequency of identifying dense inflammatory infiltrates in frozen histologic sections during Mohs operation for BCC or SCC in patients with CLL and organ-transplant recipients, to characterize the infiltrate (reactive versus leukemic) in CLL, and to estimate the subclinical tumor extension in patients with CLL, transplant recipients, and control subjects undergoing Mohs procedure. METHODS: Frozen sections of head and neck BCC and SCC obtained during Mohs procedures in patients with CLL, organ transplant recipients, and a control group were reviewed retrospectively. Biopsy specimens of CLL with dense infiltrates were assessed with immunohistochemical stains. Subclinical tumor extension (postoperative defect size minus preoperative tumor size) was evaluated in each group. RESULTS: Dense infiltrates were found in tumors of 20 of 55 patients with CLL (36%), 1 of 8 transplant recipients (13%), and 1 of 105 controls (1%). In patients with CLL, 75% of the dense infiltrates were B-cell leukemic. Compared with controls, the mean subclinical tumor extension was larger in patients with CLL (P=0.029) and in transplant recipients (P=0.55). CONCLUSION: Dense leukemic infiltrates associated with BCC or SCC in CLL may complicate pathologic interpretation of Mohs surgical histologic margins and may be associated with larger postoperative defects relative to preoperative clinical tumor appearance. In patients with CLL, as in transplant recipients, SCC seems more likely to develop than BCC.     

Invasive Cutaneous Squamous Cell Carcinoma Associated with Actinic Keratosis

BACKGROUND: It is suggested that most squamous cell carcinomas in sun-exposed areas arise from preexisting solar keratosis. Actinic keratosis is thought of as being a precursor to squamous cell carcinoma. This form of squamous cell carcinoma has been considered to be a relatively benign lesion. We report a case of invasive squamous cell carcinoma associated with actinic keratosis leading to orbit destruction and meningeal infiltration. OBJECTIVE: To demonstrate that well-differentiated tumors can act extremely aggressively with the potential toward infiltrative growth patterns. METHODS: Histologically controlled surgery along with multiple radiation therapy was performed. RESULTS: The tumor progressed inducing perineural invasion, orbit infiltration, osseous destruction, and meningeal invasion. CONCLUSION: The association of squamous cell carcinoma and actinic keratosis supports the concept of a causal relation. Excision with histologic examination of actinic keratosis seems to be useful for accurate diagnosis. Squamous cell carcinoma can represent an aggressive tumor with infiltrative growth pattern and should not be considered a benign lesion.     

Treatment of Synchronous Mucinous Carcinoma and Endocrine Mucin–Producing Sweat Gland Carcinoma with Mohs'' Micrographic Surgery

BACKGROUND: Endocrine mucin-producing sweat gland carcinoma is a very rare cutaneous tumor that has been reported only in three patients previously. We report a case of an endocrine mucin-producing sweat gland carcinoma associated with mucinous carcinoma treated by Mohs' micrographic surgery. OBJECTIVE: The purpose of this report is to test the utility of Mohs' micrographic surgery in the treatment of mucinous carcinomata. METHODS: A 79-year-old female with a 2-year history of four lesions of biopsy-proven endocrine mucin-producing sweat gland carcinomas and mucinous carcinoma was treated with Mohs' micrographic surgery. RESULTS: Three of the lesions were completely cleared by Mohs' micrographic surgery. The fourth lesion, in the right lateral canthus, was not cleared by the Mohs' technique because of its location within the orbit and the difficulty of retraction of the globe for appropriate visualization and excision. The patient underwent wide excision of the remaining orbital tumor and reconstruction, which was successfully accomplished. The patient did not experience a recurrence in any of her four lesions over a 2-year follow-up period. CONCLUSION: Mohs' micrographic surgery is an appropriate treatment for mucinous carcinomata, including endocrine mucin-producing sweat gland carcinoma and mucinous carcinoma.     

Multi-institutional long-term experience with conservative surgery for invasive penile carcinoma

PURPOSE: Invasive squamous cell carcinoma of the penis occurs on the glans, prepuce, glans and prepuce, coronal sulcus and shaft. Penile squamous cell carcinoma subsequently invades local structures, corpora cavernosa and the urethra, and metastasizes to the inguinal lymph nodes. Invasive squamous cell carcinoma of the penis usually requires total or partial penectomy. We studied the effect of primary tumor resections tailored to the anatomical extent of the cancer with preservation of uninvolved structures in select patients with invasive penile squamous cell carcinoma. MATERIALS AND METHODS: A total of 30 patients between 39 and 82 years old were treated with unconventional conservative surgical excision of the primary penile lesion. More than 130 patients were excluded from the study because they were treated with partial or total penectomy, Mohs' surgery or more extensive surgery. The 30 patients underwent preoperative biopsy with careful mapping of the extent of the disease. Patient age, tumor extent and grade, operative details, outcome and length of followup were analyzed. RESULTS: Tumor size ranged from 1.5 to 8 cm. in diameter. Tumors were well differentiated in 19 patients, moderately differentiated in 5 and poorly differentiated in 6. A total of 17 patients underwent ilioinguinal lymphadenectomy, 12 of whom had pathologically positive lymph nodes. Inguinal radiation was used in 2 patients. Chemotherapy was given to 7 patients with extensive inguinal lymphadenopathy and to 2 of 5 with pathologically positive lymph nodes. Followup ranged from 12 to 360 months. A total of 21 patients had no evidence of disease at last followup. Tumor resection with no sacrifice of function was performed in 2 patients in whom 3 small recurrences developed. One patient with numerous tumors had 2 small recurrences, which were completely excised with no further recurrence. Of the 7 patients with advanced lymphadenopathy 5 and of 5 patients with pathologically positive lymph nodes at presentation 1 died of the cancer but had no local recurrence in the penis. CONCLUSIONS: In a minority of patients with anatomically suitable penile cancer conservative surgical techniques are safe and provide equal tumor control compared to conventional resections. The anatomical situation and tumor characteristics should dictate the choice of treatment for the primary penile lesion. Inguinal lymph nodes should be managed by appropriately established guidelines but should not influence the extent of primary penile lesion resection.     

Delayed reconstruction following Mohs' chemosurgery for skin cancers of the head and neck

The case records of 52 patients with 55 cutaneous neoplasms treated by Mohs' chemosurgery and subsequently reconstructed by plastic surgeons were reviewed to determine if delay between resection and reconstruction adversely affected the outcome of reconstruction. Reconstruction was performed from 5 to 61 days after Mohs' chemosurgery for 45 basal cell carcinomas and 10 other cutaneous neoplasms. There were no complications during the interval between resection and reconstruction. Following reconstruction, minor wound complications occurred in 6% of patients; there were no major complications. Microscopic examination of the re-excised wound revealed residual disease in 2 of 45 cases of basal cell carcinoma and 0 of 10 other cutaneous malignancies. Both patients with residual basal cell carcinomas (i.e., false-negative margins after Mohs' surgery) had presented to the Mohs' surgeon with recurrent tumors. During a follow-up period of 3 months to 3 years after complete resection, recurrent tumor developed in 2 of 45 cases of basal cell carcinoma and 3 of 8 cases of squamous cell carcinoma. Delayed reconstruction, usually 5 to 20 days after Mohs' chemosurgery, can be performed without significant morbidity. Re-excision of the Mohs' chemosurgical wound for pathologic examination can detect residual disease and may be especially indicated for large recurrent wounds.     

Rim versus sagittal mandibulectomy for the treatment of squamous cell carcinoma: two types of mandibular preservation

BACKGROUND: The role of conservative mandibulectomy for patients with bone invasion from squamous cell carcinoma remains poorly defined. However, marginal mandibular resection is biomechanically secure in its design while maintaining the mandibular continuity. This procedure has proven to be a successful method of treating squamous cell carcinoma with limited mandibular involvement. PURPOSE: The purpose of this study was to analyze our results after the use of a marginal technique for the treatment of oral and oropharyngeal cancer and to compare two types of mandibular conservative procedures: rim resection versus sagittal inner mandibulectomy. METHODS: A retrospective review of a cohort of 50 patients (global group) who underwent mandibular conservative resection for previously untreated squamous cell carcinoma was performed. Two subgroups were considered: rim group (n = 37) and sagittal group (n = 13). Clinical evaluation and preoperative radiologic studies were the means used to evaluate bony invasion and to decide on the extent of mandibulectomy. The treatment outcome after these two types of mandibular resection was calculated and compared using analysis by the Pearson chi(2) test, logistic regression model for multivariate analysis, and the Kaplan-Meier method to determine survival. RESULTS.: In the sagittal group, specimens from 2 patients (11.7%) demonstrated tumor invasion on decalcified histologic examination, whereas the rim group showed 11 cases (29.7%) with bone invasion. Local recurrence was observed in the follow-up of 10 patients. No statistical relationship was found between the presence of histologic bone invasion and the risk of local recurrence. The size of bone resection >4 cm (p =.002) and tumor invasion of surgical margins (p =.039) were found to be associated with increased local recurrence rates. In multivariate analysis, lymph node affectation significantly correlated with histologic mandibular involvement (p =.02). In the global group, the 5-year observed survival rate was 56.97%. Overall survival and rate of recurrence were comparable in both groups. In the global group, tumor infiltration beyond the surgical margin was statistically related with poor survival (p =.01). CONCLUSIONS: Analysis of this series disclosed that marginal mandibulectomy is effective in the control of squamous cell carcinomas that are close to or involving the mandible. In carefully selected patients, sagittal bone resection seems to be as appropriate as rim resection in the local control of these tumors.     

Urothelial carcinoma after cyclophosphamide therapy

We report a case of transitional cell carcinoma of the bladder after cyclophosphamide therapy for chronic lymphocytic leukemia. Urologic evaluation and histologic confirmation occurred only after low grade tumor cells were identified in multiple urine samples. Malignant cells can be distinguished from drug effect in urinary specimens from patients treated with cyclophosphamide and all such cases should include regular cytologic evaluations.     

Adenoid cystic carcinoma: use of cell proliferation, BCL-2 expression, histologic grade, and clinical stage as predictors of clinical outcome

BACKGROUND: Although the three basic histologic growth patterns of adenoid cystic carcinomas (tubular, cribriform, and solid) provide some indication of clinical outcome, additional, perhaps superior, predictors of biologic activity are needed for patient management. METHODS: This series is composed of 31 adenoid cystic carcinomas that presented in Link?ping between 1982 and 1997. The tumors were clinically staged and histologically graded. For each case, after immunohistochemical identification, the proportion of tumor cells expressing the cell cycle markers MIB-1 and bcl-2 (as an indicator of proliferation and apoptosis, respectively) were quantified. Statistical correlation was sought between tumor stage and grade and the two cell cycle markers. RESULTS: The proportions of cycling tumor cells in adenoid cystic carcinomas ranged from 0.3% to 55%. For patients with no evidence of disease and a follow-up of at least 5 years, the mean percent MIB-1 value was significantly lower than for those patients who were alive with local recurrence and/or metastasis or who had died from their adenoid cystic carcinoma (p =. 024). MIB-1 tumor cell positivity also correlated strongly with tumor grade (p =.053), but not with stage (p =.22). Neither clinical stage nor histologic grade correlated with the degree of bcl-2 tumor cell positivity (p =.97 and p =.49, respectively). CONCLUSIONS: Staging and grading continue to play a vital role in the management of patients with adenoid cystic carcinoma. Furthermore, in this series of patients with adenoid cystic carcinoma, a cycling tumor cell population as measured by the MIB-1 antibody greater than 10% indicates this group as biologically more aggressive and at an increased risk for a fatal course.     

预测淋巴结阳性阴茎癌手术后无病生存率列线图的建立

目的 建立预测淋巴结阳性阴茎癌患者手术后无病生存率的列线图. 方法 回顾性分析32例淋巴结阳性阴茎鳞状细胞癌患者的临床病理资料.患者均接受了阴茎肿瘤切除和区域淋巴结清扫术.N1期10例,N2期13例,N3期9例.术前鳞状细胞癌抗原(SCCAg)水平0～19.8μg/L,中位值为0.9μg/L.随访3～31个月,中位时间16.5个月.患者出现疾病复发13例.预后指标包括分期、分级和术前SCCAg水平.采用Cox回归模型建立列线图. 结果 血清SCCAg水平的风险比值为2.76(95％ CI:1.40～5.44,P=0.0034),而N分期的风险比值为28.51 (95％ CI:2.55～319.11,P=0.0066).列线图一致系数为0.862,经过白助法校正后为0.855;校正曲线显示模型的符合度良好.N分期和术前SCCAg水平是预测淋巴结阳性阴茎癌患者手术后无病一存率的独立预后因素. 结论 基于分期和术前SCCAg水平,构建预测淋巴结阳性阴茎癌手术后无病生存率的列线图,有助于个体化判断肿瘤复发的风险及与患者交流和综合治疗选择.     

Curettage prior to Mohs'' Micrographic Surgery for Previously Biopsied Nonmelanoma Skin Cancers: What Are We Curetting? Retrospective, Prospective, and Comparative Study

BACKGROUND: Curettage prior to excision and Mohs' micrographic surgery for nonmelanoma skin cancer is performed based on the assumption that the curette will remove softer, more friable tumor-infiltrated dermis and leave structurally intact normal skin. This assumption, however, has not been objectively examined in the dermatologic surgery literature. OBJECTIVE: We performed a study to examine the ability of curettage to selectively remove and delineate nonmelanoma skin cancer prior to Mohs' micrographic surgery. METHODS: The study included 150 previously biopsied basal cell and squamous cell carcinomas less than 1.5 cm in size. We conducted (1) a retrospective study of 50 tumors curetted prior to Mohs' surgery by a surgeon who routinely curettes preoperatively; (2) a prospective study in which a surgeon who routinely does not curette preoperatively curetted 50 tumors prior to Mohs' surgery; and (3) a comparative historical group of 50 noncuretted tumors treated with Mohs' surgery by the latter surgeon. All curetted tissue was evaluated histologically. RESULTS: Only 50% of the curetted tissue demonstrated the presence of tumor in the curettings, but in 76% of these, the curette left residual tumor at the surgical margins. Of the other 50% in which the curette removed only non-cancer-containing skin, 34% had tumor present at the surgical margin. Overall, the curette removed tumor, leaving no residual tumor at the surgical margins in only 12% of lesions. Comparison with historical noncuretted tumors operated on by the same surgeon showed that curettage did not affect the mean number of stages or the proportion of tumors requiring more than one stage for histologic clearance. CONCLUSION: Although curettage may be helpful in debulking friable skin prior to Mohs' micrographic surgery, it does not reliably delineate the extent of a tumor.     

Indication for limited surgery on small lung cancer tumors measuring 1cm or less in diameter on preoperative computed tomography and long-term results

OBJECTIVE: Non-invasive lung cancers showed a good prognosis after limited surgery. But it is still uncertain about invasive lung cancers. We investigated the indications for limited surgery for small lung cancer tumors measuring 1 cm or less in diameter on preoperative computed tomography (CT). METHODS: This study retrospectively analyzed of 1,245 patients who underwent complete resection of lung cancer between 1989 and 2004 in our hospital. Sixty-two patients (5%) had tumors measuring 1 cm or less in diameter. The probability of survival was calculated using the Kaplan-Meier method. RESULTS: All diseases were detected by medical checkup, 52 % of the patients were not definitively diagnosed with lung cancer before surgery. Adenocarcinoma was histologically diagnosed in 49 patients (79%). Other histologic types included squamous cell carcinoma (8), large cell carcinoma (1), small cell carcinoma (1), carcinoid (2), and adenosquamous cell carcinoma (1). Fifty-seven patients (92%) showed pathologic stage IA. The other stages were IB (2), IIA (1), and IIIB (2). There were 14 bronchioloalveolar carcinomas (25% of IA diseases). The 5-year survival rates of IA patients were 90%. The 5-year survival rate of patients with tumors measuring 1cm or less diameter was 91% after lobectomy or pneumonectomy, and 90% after wedge resection or segmentectomy. There were 3 deaths from cancer recurrence, while there were no deaths in 14 patients with bronchioloalveolar carcinoma CONCLUSION: After limited surgery, non-invasive cancer showed good long-term results, while invasive cancer showed a recurrence rate of 2.3% to 79% even though the tumor measured 1 cm or less in diameter on preoperative CT.     

Mucoepidermoid and squamous cell carcinomas of breast with reference to squamous metaplasia and giant cell tumors

Clinical and pathologic features of five low-grade mucoepidermoid and a similar number of squamous cell carcinomas of breast are presented. Three of each were retrieved from 55 cases of invasive breast cancer previously assessed to exhibit squamous metaplasia. Although the estimated incidence of low-grade mucoepidermoid cancers is approximately 0.2%, they may be more frequent, masquerading under designations of squamous metaplasia or intracystic carcinomas. All patients with low-grade mucoepidermoid cancer, and the two recorded previously, were free of recurrence for 4-10 years. No high-grade mucoepidermoid cancers were found. Squamous cell cancers frequently have a phyllode configuration (carcinoma phyllodes). Two examined by electron microscopy revealed a commonality of ultrastructural features with similar tumors of other sites. None exhibited glandular differentiation or tinctorial evidence of mucin secretion. One squamous cell carcinoma contained both malignant epithelial and banal, osteoclast-like, stromal giant cells. Only the latter were identified in tissue culture. This experience and review of the literature prompted a nosologic characterization of giant cell lesions of breast. One patient with squamous cell carcinoma succumbed because of her disease; the remainder are either free of or alive with recurrence at 4 and 10 years.     

A rationale for the treatment of difficult basal cell and squamous cell carcinomas of the skin

During a 56-month period, 1,348 patients with either basal cell or squamous cell carcinoma of the skin underwent microscopically controlled excision (Mohs' fresh-tissue technique) by a dermatologic surgeon. Of these 1,348 patients, 394 required various types of wound reconstruction by a plastic surgeon. The average dimensions of these wounds were 5 X 3.5 cm, with a range of 1.5 to 20.0 cm in diameter. There were postoperative complications in 13.9% of the 394 patients. The recurrence rate in these 394 patients, to date, for basal cell carcinoma is 5.2% and for squamous cell carcinoma, 11.9%. The advantages of this treatment arrangement include maximum conservation of normal tissue with provision of more reliable complete tumor excision, optimum time utilization by both dermatologic and plastic surgeons, separation of the responsibility for tumor ablation from that for wound reconstruction, and, in many instances, economic savings. We recommend this method of management for difficult, high-risk basal cell or squamous cell carcinomas.     

Clinicopathologic characteristics of peripheral squamous cell carcinoma of the lung

Squamous cell carcinoma of the lung can be divided into two types according to the location of the primary site: the central type and the peripheral type. The clinicopathologic factors in the peripheral type of lung squamous cell carcinoma have not yet been fully evaluated. A total of 204 surgically resected lung squamous cell carcinomas were reviewed with special reference to their location, histologic characteristics based on tumor growth patterns, and clinicopathologic factors. The central type and the peripheral type accounted for 95 and 109 cases, respectively. Although the patient population of the peripheral type was older, with a lower pathologic stage, lower lymphatic vessel involvement, and lymph node metastasis, the Kaplan-Meier survival proportions did not differ significantly between these two groups. Based on the histologic growth pattern, the peripheral type was classified under three subgroups as follows: 1). the alveolar space-filling type, 2). the expanding type, and 3). the combined type. Among these three types, the alveolar space-filling type showed neither lymphatic vessel invasion nor lymph node metastasis and had the most favorable prognosis. The central and peripheral types of lung squamous cell carcinoma have different clinicopathologic characteristics and should be classified under respectively different categories.     

Angiosarcoma Arising from the Tongue of an 11-Year-Old Girl with Xeroderma Pigmentosum

Xeroderma pigmentosum (XP) is a rare autosomal recessive defect in DNA endonuclease activity that is associated with the development of cutaneous malignancies, at sun exposed sites, including basal cell carcinoma, squamous cell carcinoma, and melanoma. Squamous cell carcinomas are also known to target the anterior tongue. Patients sometimes develop angiosarcomas, and these invariably arise from sun-exposed skin. A biopsy was taken from a large mass arising in the anterior tongue of an 11-year-old girl with XP and a history of cutaneous basal cell carcinomas. The histopathologic findings demonstrated a high grade epithelioid neoplasm resembling a poorly differentiated squamous cell carcinoma, but the immunohistochemical profile (AE1/AE3 negative, p63 negative, CD31 positive, CD34 positive) established the diagnosis of angiosarcoma. Angiosarcoma is an XP-related tumor that usually arises in sun-exposed skin but can also arise in the oral cavity. For patients with XP who develop epithelioid neoplasms of the oral cavity, epithelioid angiosarcoma should be considered in the differential diagnosis.     

Squamous cell carcinoma of the perineum: management with Mohs' surgical excision and gluteal thigh flap

Two patients with squamous cell carcinoma of the perineum are presented. The carcinomas were treated by wide local excision using Mohs' fresh-tissue technique for control of the surgical margins. The wounds were closed using a gluteal thigh flap. In 1 patient the flap was extended across the midline for coverage of a contralateral perineal site. Healing was complete in each case. The combination of Mohs' fresh-tissue methods and the posterior thigh flap provided accurate identification and removal of the perineal tumor with stable wound coverage.     

Malignant salivary tumors—analysis of prognostic factors and survival

A group of 113 patients with malignant salivary gland tumors was retrospectively reviewed to analyze the association of clinical and histologic factors with survival. These factors were patient sex and age, tumor site, clinical stage, histologic diagnosis, tumor grade, and whether or not final surgical margins were clear. There were 57 parotid, 40 minor salivary, and 16 submandibular gland cancers. The histologic groups were mucoepidermoid carcinoma (49 patients), adenoid cystic carcinoma (31), adenocarcinoma not otherwise specified (18), acinic cell carcinoma (7), malignant mixed tumor (5), squamous cell carcinoma (2), and undifferentiated carcinoma (1). Univariate analysis of clinical factors showed that age and clinical stage significantly influenced survival. At 10 yr the predicted cumulative survival rates for Stage I, II, III, and IV tumors were 74%, 56%, 32%, and 10%, respectively. Tumor grade was the only significant histologic factor. This was most obviously reflected among patients with mucoepidermoid carcinomas. Cumulative survival at 5 yr was 94% for those with low-grade tumors and 26% for high-grade tumors. By multivariate analysis, clinical stage, age, and tumor grade remained highly significant. Analysis of patients with only Stage I and II disease demonstrated that the significant factors were patient age, tumor site, tumor grade, and whether or not surgical clearance was achieved. These results suggest that clinical stage should not be the exclusive determinant of the extent of surgery and that the selection of patients, for adjuvant therpay may be improved by an awareness of these prognostic factors.     

Malignant salivary tumors--analysis of prognostic factors and survival

A group of 113 patients with malignant salivary gland tumors was retrospectively reviewed to analyze the association of clinical and histologic factors with survival. These factors were patient sex and age, tumor site, clinical stage, histologic diagnosis, tumor grade, and whether or not final surgical margins were clear. There were 57 parotid, 40 minor salivary, and 16 submandibular gland cancers. The histologic groups were mucoepidermoid carcinoma (49 patients), adenoid cystic carcinoma (31), adenocarcinoma not otherwise specified (18), acinic cell carcinoma (7), malignant mixed tumor (5), squamous cell carcinoma (2), and undifferentiated carcinoma (1). Univariate analysis of clinical factors showed that age and clinical stage significantly influenced survival. At 10 yr the predicted cumulative survival rates for Stage I, II, III, and IV tumors were 74%, 56%, 32%, and 10%, respectively. Tumor grade was the only significant histologic factor. This was most obviously reflected among patients with mucoepidermoid carcinomas. Cumulative survival at 5 yr was 94% for those with low-grade tumors and 26% for high-grade tumors. By multivariate analysis, clinical stage, age, and tumor grade remained highly significant. Analysis of patients with only Stage I and II disease demonstrated that the significant factors were patient age, tumor site, tumor grade, and whether or not surgical clearance was achieved. These results suggest that clinical stage should not be the exclusive determinant of the extent of surgery and that the selection of patients, for adjuvant therapy may be improved by an awareness of these prognostic factors.