Search

Background. The efficacy of focal continuous low dose-rate brachytherapy (CLDR-BT) for prostate cancer requires that appropriate margins are applied to ensure robust target coverage. In this study we propose a method to establish such margins by emulating a focal treatment in patients treated with CLDR-BT to the entire gland.Material and methods. In 15 patients with localized prostate cancer, prostate volumes and dominant intra-prostatic lesions were delineated on pre-treatment magnetic resonance imaging (MRI). Delineations and MRI were registered to trans-rectal ultrasound images in the operating theater. The patients received CLDR-BT treatment to the total prostate volume. The implantation consisted of two parts: an experimental focal plan covering the dominant intra-prostatic lesion (F-GTV), followed by a plan containing additional seeds to achieve entire prostate coverage. Isodose surfaces were reconstructed using follow-up computed tomography (CT). The focal dose was emulated by reconstructing seeds from the focal plan only. The distance to agreement between planned and delivered isodose surfaces and F-GTV coverage was determined to calculate the margin required for robust treatment.Results. If patients had been treated only focally, the target volume would have been reduced from an average of 40.9 cm3 for the entire prostate to 5.8 cm3 for the focal plan. The D90 for the F-GTV in the focal plan was 195 ± 60 Gy, the V100 was 94% [range 71–100%]. The maximum distance (cd95) between the planned and delivered isodose contours was 0.48 cm.Conclusions. This study provides an estimate of 0.5 cm for the margin required for robust coverage of a focal target volume prior to actually implementing a focal treatment protocol.

Background and purpose The reliability of post-implant dosimetry in the OR depends on the geometrical variability of implant and anatomy after the procedure. The purpose was to gain detailed information on seed displacement patterns in different sectors of the prostate. Materials and methods Of 33 patients with stranded seed implants the seed geometry and the dose distribution were compared between the situation in the OR just after the procedure, based on ultrasound images, and the situation after 1month, based on registered CT and MR images. Results There was a substantial displacement of ventral seeds of 3.8±2.5mm in caudal direction ( p p D 90 in the dorsal-caudal and ventral-caudal sectors increased with respectively 44±20Gy and 29±28Gy ( p p =0.008). Conclusions There were substantial changes in dose distribution 1month after the procedure, mainly due to implant and prostate shrinkage and displacement of ventral seed strands in caudal direction. When performing dynamic dosimetry or dosimetry at the end of the procedure the effect of these phenomena has to be taken into account when using stranded seeds.

Background and Purpose To investigate possible relationships between the dose to the sub-segments of the lower urinary tract and lower urinary tract symptoms (LUTS) after brachytherapy of the prostate. Materials and Methods This study involved 225 patients treated for prostate cancer with I-125 seeds. Post-implant dose-volume histograms of the prostate, urethra, bladder wall, bladder neck and external sphincter were determined. Endpoints were the mean and the maximum International Prostate Symptom Score (IPSS) during the first 3 months after the treatment. For binary analysis the patients were stratified in a group with enhanced LUTS and a group with non-enhanced LUTS. Results The dose to 0.5 cm 3 of the bladder neck ‘ D 0.5cc -blne’ ( p = 0.002 and p = 0.005), the prostate volume prior to treatment ‘ V pr -0’ ( p = 0.005 and p = 0.024) and the pre-treatment IPSS (both p D 0.5cc -blne ⩾ 175 Gy and a V pr -0 ⩾ 42 cm 3 , 68% suffered from enhanced LUTS, against just 30% of the other patients ( p Conclusions Pre-treatment IPSS, prostate volume and dose to the bladder neck are correlated with post-implant IPSS. A combination of a large prostate and a high dose to the bladder neck is highly predictive for enhanced early LUTS.

We assessed the prognostic significance of extranodal extension, defined as tumor extension through the lymph node capsule into the perinodal fibrous-adipose tissue, as well as several other risk factors in node positive penile cancer cases.We analyzed prospectively collected data on a consecutive series of 156 chemotherapy naïve patients with proven lymph node involvement who underwent therapeutic regional lymphadenectomy. Postoperative external radiotherapy was indicated when histopathological analysis revealed more tumor than 1 intranodal metastasis. We estimated cancer specific survival using the Kaplan-Meier method. Multivariate analysis was done according to the Cox proportional hazards model of factors statistically significant on univariate analysis.Adjuvant radiotherapy was done in 70 patients (45%). Median followup was 57.8 months. Overall 5-year cancer specific survival was 61%. Men with extranodal extension had significantly decreased 5-year cancer specific survival compared with men without it (42% vs 80%). Other prognostic variables on univariate analysis were bilateral metastatic involvement vs unilateral, 3 or greater unilateral metastatic inguinal nodes vs 2 or fewer, inguinal lymphadenectomy positive margin status vs negative status and pelvic lymph node involvement. Pathological T stage or differentiation grade were not significant predictors of outcome. On multivariate analysis extranodal extension and pelvic lymph node involvement remained associated with decreased cancer specific survival (HR 2.37 and 2.20, respectively).Metastatic inguinal lymph node extranodal extension and pelvic lymph node involvement are independent predictive parameters of cancer specific survival in patients with pathologically node positive penile carcinoma despite surgery with postoperative radiotherapy.

Purpose Reduction of the number of implantation needles for prostate brachytherapy will shorten the duration of implantation procedures and possibly reduce trauma-related morbidity. The purpose of this study was to investigate possibilities for the minimization of the number of needles and to investigate the consequences for the dose distribution. Methods and Materials A planning study for six different prostate volumes was performed. The number of needles was minimized by changing fixed 1 cm interseed spacing to free interseed spacing within the needles and by increasing the seed activity. Dose–volume parameters of prostate and organs at risk (OAR) bladder, rectum, and urethra were determined. For plans with different needle and seed configurations, the sensitivity for random seed placement inaccuracies was tested. Dose distributions of realized implants based on fixed ( n = 21) and free interseed spacing (n = 21) were compared. Results The average number of needles (±1 standard deviation) could be reduced from 18.8 ± 3.6 to 12.7 ± 2.9 (−33%) when changing from fixed interseed spacing to free interseed spacing and subsequently to 7.3 ± 1.0 (−42%) by increasing the seed strength from 0.57 U to 1.14 U. These needle reductions resulted in increased dose inhomogeneity within the prostate and increased sensitivity of dose–volume parameters of the OAR for random geometrical inaccuracies. Introduction of free interseed spacing in our clinic resulted in very satisfactory dose coverage of the prostate ( D 90 = 172 ± 17 Gy), while the average number of needles was reduced by 30%. Conclusions Substantial reduction of the number of implantation needles is possible without compromising adequate dose coverage of the prostate. However, the chance of an unpredicted high dose to the OAR increases as fewer needles are used.

Purpose After prostate implantation, dose calculation is usually based on a single imaging session, assuming no geometrical changes occur during the months of dose accumulation. In this study, the effect of changes in anatomy and implant geometry on the dose distribution was investigated. Materials and methods One day, 1 month and 3 1 2 months after seed implantation, a combined TRUS-CT scan was made of 13 patients. Based on these scans changes in dose rate distribution were determined in prostate, urethra and bladder and a ‘geometry corrected' dose distribution was estimated. Results When based on the day-1 scan, parameters representing high dose volumes in prostate and urethra were largely underestimated: V150 of the prostate 18±10% and V120 of the urethra 47±32%. The dose to a 2cm 3 hotspot in the bladder wall (D2cc), however, was overestimated by 31±35%. Parameters based on scans 1 month post-implant or later were all within ±5% of geometry corrected values. Conclusion Values meant to indicate the adequacy of dose coverage of the prostate, V100 and D90, were not influenced by geometrical changes and were independent of the post-implant scan date. Other parameters representing high dose volumes changed strongly within the first month after implantation.

Objective: To evaluate the long-term survival following brachytherapy and following cystectomy of patients with invasive bladder cancer treated in our institution. Patients and methods: Between 1988 and 2000 108 patients with solitary, organ confined T1–T2 invasive bladder cancer of ≤5cm were treated with a transurethral resection, and a course of external beam radiotherapy (30Gy) followed by 40Gy brachytherapy. The overall and disease specific survival rates of these patients are compared with those of 77 patients with T1–T2 invasive bladder cancer treated with cystectomy between 1988–2003. Results: The 5/10 year overall survival rates were 62%/50% after brachytherapy and 67%/58% after cystectomy ( p =0.67). The 5/10 year disease specific survival rates were 73%/67% after brachytherapy and 72%/72% after cystectomy ( p =0.28). When adjusted for age, multiplicity, T-stage, N-stage and grade, the 5/10 year overall survival rates were 65%/53% after brachytherapy and 62%/51% after cystectomy, respectively. The adjusted disease specific survival rates were 75%/70% after brachytherapy and 66%/66% after cystectomy. Conclusions: This study does not provide evidence regarding survival against the use of bladder preservation with brachytherapy for patients with solitary, T1–T2 invasive bladder cancer of ≤5cm diameter, seeking bladder-sparing alternatives to radical cystectomy.

Purpose To evaluate our long-term results of bladder preservation with brachytherapy in the treatment of bladder cancer. Methods and materials Between 1987 and 2000, 108 patients with T1-G3 and T2-T3a stages of bladder cancer were treated with a transurethral resection (TUR) and a course of external beam radiotherapy (30 Gy in 15 fractions) followed by brachytherapy (40 Gy). All tumors were solitary lesions with a diameter ≤5 cm. Median follow-up was 54 months (range, 1–178 months). Results The 5-year and 10-year overall survival rates were 62% and 50%, respectively. The 5-year and 10-year disease-specific survival rates were 73% and 67%, respectively. The actuarial local control rate was 73% at 5 and 73% at 10 years, respectively. The 5-year and 10-year disease-specific survival rates for patients with a preserved bladder were 68% and 59%, respectively. Of all long-term surviving patients, 90% preserved their native bladders. The treatment was well tolerated. Acute toxicity was mild. Two patients experienced serious late toxicity: 1 patient developed a persisting vesicocutaneous fistula and the other a stricture of the urethra and ureters. Conclusion For patients with solitary, organ confined invasive bladder cancer ≤5 cm, bladder preservation with brachytherapy is an excellent alternative to radical cystectomy and combined modality treatment.

Background and purpose We investigated the application of a differential target- and dose prescription concept for low-dose-rate prostate brachytherapy (LDR-BT), involving a re-distribution of dose according to risk of local failure and treatment-related morbidity. Material and methods Our study included 15 patients. Multi-parametric MRI was acquired prior to LDR-BT for gross tumor volume (GTV) delineation. Trans-rectal ultrasound (US) images were acquired during LDR-BT for prostate gland- (CTV Prostate ) and organs at risk delineation. The GTV contour was transferred to US images after US/MRI registration. An intermediate-risk target volume (CTV Prostate ) and a high-risk target volume (CTV HR =GTV+5mm margin) were defined. Two virtual dose plans were made: Plan risk-adapt consisted of a de-escalated dose of minimum 125Gy to the CTV Prostate and an escalated dose to 145–250Gy to the CTV HR ; Plan ref included the standard clinical dose of minimum 145Gy to the CTV Prostate . Dose-volume-histogram (DVH) parameters were expressed in equivalent 2Gy fractionation doses. Results The median D 90% to the GTV and CTV HR significantly increased by 44Gy and 17Gy, respectively when comparing Plan risk-adapt to Plan ref . The median D 10% and D 30% to the urethra significantly decreased by 9Gy and 11Gy, respectively and for bladder neck by 18Gy and 15Gy, respectively. The median rectal D 2.0cm 3 had a significant decrease of4Gy, while the median rectal D 0.1cm 3 showed an increase of 1Gy. Conclusions Our risk adaptive target- and dose prescription concept of prescribing a lower dose to the whole gland and an escalated dose to the GTV using LDR-BT seed planning was technically feasible and resulted in a significant dose-reduction to urethra and bladder neck.

Background and purpose : The three-dimensional (3D) dose distribution in combination with 3D anatomy of 13 patients treated for cervical carcinoma with intracavitary brachytherapy was analyzed. The aim of this study was to determine the correlation between a dose value obtained from the integral dose volume histogram (DVH) of the rectum and (a) the Nederlands Kanker Instituut (NKI) point of reference for the rectum dose ( R ) and (b) the highest dose to the frontal rectum wall in the transverse CT slice near the top of the vagina through point R . Results : The correlation between the DVH rectum dose value for 2 cm 3 in the highest dose region and the rectum dose at point R was poor (regression coefficient 0.50). On the contrary, however, the correlation between the DVH rectum dose value for 2 cm 3 in the highest dose region and the maximum rectum dose value in a transverse CT slice through point R was good (regression coefficient 0.90). Conclusions : The maximal rectum dose value obtained from a transverse CT slice near the top of the vagina through point R was found to be a more representative point for the rectal dose burden and might therefore show a good correlation with complications. The point of reference for the rectal dose ( R ) was found not to be a reliable estimation of the maximal dose in the rectum.

Purpose: A disadvantage of ovoid shields in a Fletcher-type applicator is that these shields cause artifacts on postimplant CT images. CT images, however, make it possible to calculate the dose distribution in the rectum and the bladder. To be able to estimate the possible advantage of having CT information over the use of ovoid shields without having CT information, we investigated the influence of shielding segments in a Fletcher-type Selectron-LDR applicator on the dose distribution in rectum and bladder. Methods and Materials: Contours of rectum and bladder were delineated on transaxial CT slices of 15 unshielded applications the volumes contained within these structures dose-volume histograms (DVHs) were calculated. In a similar way, DVHs of simulated shielded applications were calculated. The reduction, due to shielding, of the dose to the 2 cm 3 (D 2 ) and 5 cm 3 (D 5 ) volume of the cumulative DVHs of rectum and bladder, were determined. An isodose pattern in the sagittal plane through the center of each applicator was plotted to compare the location of the shielded area with the location of maximum dose in rectum and bladder in the unshielded situation. In two cases local dose reductions to the rectal wall were determined by calculating the dose in points at 10-mm intervals on the rectal contours. Results: For the rectum, the reduction of D 2 ranged from 0 to 11.1%, with an average of 5.0%; the reduction of D 5 ranged from 2.3 to 12.1%, with an average of 6.4%. The reduction of D 2 and D 5 for the bladder ranged from 0 to 11.9% and from 0 to 11.6%, with average values of 2.2 and 2.6%, respectively. In 8 out of 15 eases the rectal maximum dose was located inferior to the shielded area. In all cases except one the bladder maximum dose was located superior to the shielded area. Local dose reductions on the rectal wall can be as high as 30% or more in an optimally shielded area. Conclusions: Reductions of D 2 and D 5 to rectum and bladder due to shielding are rather small, because the shielded area does usually not coincide with the high dose region and even if it does, the shielded area is too small to result in large reductions of these values. Because local dose reductions vary largely, one should proceed with caution when calculating the dose in just one rectal or bladder reference point. Because large overall dose reductions cannot be achieved with shielding, it is safe to use an unshielded applicator when post implant CT images are used to realize optimized dose distributions.

Squamous cell carcinoma of the penis is a rare disease in developed countries. Based on epidemiological data from the comprehensive cancer centers in the Netherlands, the age-adjusted incidence of penile cancer is estimated to be 1.5/100,000. Despite the fact that this disease is readily visible and easily accessible for both diagnostic and surgical interventions patients still present with advanced disease.

Among 118 patients with squamous cell carcinoma of the penis treated at our cancer institute between 1956 and 1989, we analyzed the accuracy of classification, using the tumor, nodes and metastasis system. We analyzed the role of lymphography, computerized tomography and fine needle aspiration cytology as additional staging procedures. The primary tumor (T category) was classified incorrectly in 26% of the cases. Overstaging was noted in 10% of the cases because of unsuspected infiltration and overstaging was noted in 16%. Overstaging occurred because of edema and infection masking the actual size and giving a misconception of infiltration, and also because of primary presentation as large exophytic tumors with no or minimal histopathological infiltration. When the regional lymph nodes were categorized simply as positive or negative 80% of the tumors were classified correctly and 20% incorrectly (13% were false positive and 7% were false negative). Regional lymph node invasion that escaped clinical examination was not detected by any imaging examination or fine needle aspiration cytology study. Positive findings were found only in patients with clinically suspected nodes. The classification of regional nodes by clinical examination only is hardly improved by additional imaging studies. Clinical decisions with respect to the management of regional lymph nodes should not be based on negative findings of lymphangiography, computerized tomography or fine needle aspiration cytology. In patients with proved metastasis additional imaging may be of some help in the detection of pelvic node invasion and the determination of the extent of involvement. We recommend lymphangiography as the examination of choice.

Purpose Lower urinary tract symptoms are frequently observed after I-125 seed implantation of the prostate. More knowledge about causes and predictors is necessary to be able to develop less toxic implantation techniques. The aim of this study was to identify implantation related factors that contribute to post-implant urinary morbidity. Materials and methods Analysed was a group of 72 patients that filled in a symptom score questionnaire before, 3 months and 6 months after implantation as well as a group of 15 patients that suffered from acute urinary retention. Several dose–volume parameters of prostate, urethra and bladder wall were determined based on a post-implant TRUS-CT scan. Results The dose to a 1cm 3 hotspot in the bladder wall (D1cc-bl) as well as the prostate volume were independently correlated with urinary morbidity symptom scores at 3 months ( p =0.006 and p =0.005, respectively) and at 6 months ( p =0.001 and p =0.015, respectively) after implantation. The number of implanted seeds and the D1cc-bl were significant discriminators ( p p =0.015, respectively) for either mild or severe early urinary morbidity. Conclusion Bladder hotspot dose appears to be an important dosimetric predictor for urinary morbidity both at 3 months and at 6 months after implantation. Other predictors are prostate volume, or equivalently, the number of implanted seeds.

We identified pathological parameters of inguinal lymph node involvement with the aim of predicting pelvic lymph node involvement and survival.A total of 308 patients with penile carcinoma and adequate followup were included in this study. The outcome of 102 patients who underwent lymphadenectomy for lymph node metastases was analyzed further. Histopathological characteristics of the regional lymph nodes were reviewed including unilateral or bilateral involvement, the number of involved nodes, pathological tumor grade of the involved nodes, and the presence of extracapsular growth.Tumor grade of the involved inguinal lymph nodes (OR 6.0, 95% CI 1.2-30.3) and the number of involved nodes (2 or less vs more than 2) (OR 12.1, 95% CI 3.0-48.1) were independent prognostic factors for pelvic lymph node involvement. Extracapsular growth (OR 2.3, 95% CI 1.1-4.8), bilateral inguinal involvement OR 3.4, 95% CI 1.2-9.4) and pelvic lymph node involvement (OR 3.1, 95% CI 1.4-6.6) were independent prognostic factors for disease specific survival.Patients with only 1 or 2 inguinal lymph nodes involved without extracapsular growth and no poorly differentiated tumor within these nodes are at low risk of pelvic lymph node involvement and have a good prognosis with a 5-year survival rate of approximately 90%. Pelvic lymph node dissection seems to be unnecessary in these cases.

Eleven patients, ranging in age from 58 to 79 years and referred for 125I implantation because of localised prostatic carcinoma, were subjected to iliopelvic lymphoscintigraphy (IPL) prior to lymph node dissection. Scintigraphy was performed by perianal injection into the ischiorectal fossa of 0.2 ml 99mTc radio-labelled antimony sulphide colloid. The sensitivity of IPL was 100% and the specificity 13%, with a prevalence of positive nodes of 27%. It was concluded that IPL is not useful as a non-surgical method of detecting lymph node invasion in localised prostatic carcinoma because of the low specificity.

Purpose: To evaluate and eventually quantify a possible influence of tumor proliferation during the external radiation course on local control in muscle invasive bladder cancer. %Methods and Materials: The influence of total dose, overall treatment time, and treatment interruption has retrospectively been analyzed in a series of 379 patients with nonmetastasized, muscle-invasive transitional cell carcinoma of the urinary bladder. All patients received external beam radiotherapy at the Netherlands Cancer Institute between 1977 and 1990. Total dose varied between 50 and 75 Gy with a mean of 60.5 Gy and a median of 60.4 Gy. Overall treatment time varied between 20 and 270 days with a mean of 49 days and a median of 41 days. Number of fractions varied between 17 and 36 with a mean of 27 and a median of 26. Two hundred and fourty-four patients had a continuous radiation course, whereas 135 had an intended split course or an unintended treatment interruption. Median follow-up was 22 months for all patients and 82 months for the 30 patients still alive at last follow-up. A stepwise procedure using proportional hazard regression has been used to identify prognostic treatment factors with respect to local recurrence as sole first recurrence. Results: One hundred and thirty-six patients experienced a local recurrence and 120 of these occurred before regional or distant metastases. The actuarial local control rate was 40.3% at 5 years and 32.3% at 10 years. In a multivariate analysis total dose showed a significant association with local control ( p = 0.0039), however in a markedly nonlinear way. In fact only those patients treated with a dose below 57.5 Gy had a significant higher bladder relapse rate, whereas no difference in relapse rate was found among patients treated with doses above 57.5 Gy. This remained the case even after adjustment for overall treatment time and all significant tumor and patient characteristics. The Normalized Tumor Dose (NTD) (α/β = 10) and NTD (α/β = 15) were not significantly related to local control ( p = 0.96 and p = 0.053, respectively). Only weak evidence was found for an association between local control and overall treatment time ( p = 0.067). No difference in bladder relapse rate was found among patients treated with a continuous course and patients who had treatment interruptions ( p = 0.099). Neither the length of the interruption, nor the actual number of treatment days has a significant influence on local control ( p = 0.04 and p = 0.09, respectively). Conclusion: In contrast to two earlier, but smaller reports, in this study no significant effect of treatment prolongation on outcome after radiotherapy could be demonstrated and thus no support was found for an important role for tumor proliferation as the cause of treatment failure in muscle-invasive bladder cancer. Results of large-sized phase III trials will have to be awaited to show any benefit from reduction of the overall treatment time and to quantify the potential effect of tumor proliferation.1998 Elsevier Science Inc.

Purpose: To establish the shortest possible overall treatment time, with respect to acute and late toxicity, in which a radiation dose of 66 Gy in 2 Gy/fraction can be administered to the urinary bladder. Methods and Materials: In a stepwise fashion the overall treatment time of our routinely used conventional treatment schedule (66 Gy in 2 Gy/fraction in 6.5 weeks) was reduced to 5 weeks (Step 1) and then 4 weeks (Step 2). Shortening of the overall treatment time was obtained by giving two daily fractions during the last part of the treatment. The interfraction interval was a minimum of 6 h. Toxicity was scored according to a modified scheme of the Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer. Fifteen patients received the 5-week schedule and 25 patients received the 4-week schedule. Results: Note that feasibility was very good for the 5-week schedule. No Grade ≥3 acute bowel or bladder toxicity was seen in these patients. Severe late complications did not occur. In the 4-week schedule, acute bladder toxicity was unaffected but acute bowel toxicity was considerably increased. Of the 25 patients, 16 suffered from Grade 3 toxicity, 3 from Grade 4, and 1 from Grade 5 toxicity. Actuarial analysis at 3 years showed a 31% incidence of Grade 4 late bladder complications. Conclusion: It is concluded that the 5-week schedule is well tolerated. In contrast, the 4-week schedule shows unacceptable acute bowel toxicity and an increased incidence of late bladder comlications ( p = 0.043).

OBJECTIVE To retrospectively analyse the efficacy and toxicity of conservative treatment for T1G3 and T2-3a bladder carcinoma and to compare the results with those obtained in a previous study using caesium implantation. PATIENTS AND METHODS Between 1987 and 1990 12 patients with high-grade T1 tumours and 28 with T2-3a tumours were treated using this combined approach. All tumours were solitary with a diameter not exceeding 5 cm. Treatment consisted of transurethral resection, a course of external irradiation (30 Gy in 15 fractions) to the whole pelvis and an implant procedure. RESULTS With a mean follow-up of 40 months (range 24-65) 31 patients remained free of disease and nine relapsed: three with distant metastases, two with bladder recurrence and four with combined bladder and distant relapse. Three patients showed tumour at the original site. The 5 years actuarial rate of local control was 84% with an overall 5 years actuarial survival of 86%. Early problems included bladder leakage, wound infection and psychological problems. Late complications in four patients, was transient ulceration at the implant side causing mild dysuria. CONCLUSIONS Implantation with iridium is equally effective as with caesium but has significantly fewer complications.

We analyzed the management of regional lymph nodes in 110 patients with squamous cell carcinoma of the penis treated at the Netherlands Cancer Institute between 1956 and 1989 with curative intent. Of 66 patients who presented with unsuspected nodes 57 were placed on a surveillance program, while lymph node dissection was performed in 5 (with adjuvant external radiation therapy in 1) and 4 were treated with external radiation therapy only. The management of 40 patients with clinically suspected nodes included surveillance in 5, lymph node dissection in 27 (with adjuvant radiotherapy in 11), biopsy in 4 and external radiation therapy in 4. Postoperative radiotherapy had been given if more than 2 nodes were involved or when extracapsular growth was observed. Overall, 25 patients had a regional recurrence, 5 of whom could be cured subsequently. All regional recurrences developed within 2 years after primary treatment. Analysis showed 100% survival in histologically proved node negative patients (stage pN0). The success of lymph node dissection was related to the extent of the metastatic spread and to the number of involved nodes. Patients with 1 positive node and unilateral inguinal involvement showed a statistically significant survival advantage compared to patients with more extensive spread. Considering the indications for node dissection we found a clear relationship among T category, grade and the probability of lymph node invasion. Patients with stage T1 tumors and stage T2, grades 1 and 2 tumors presented significantly less often with lymphatic invasion than those with other categories of disease and were less likely to have a regional recurrence after treatment of the primary tumor only. In these categories we recommend surveillance of the regional lymph nodes in patients who present with unsuspected nodes. However, patients with stage T2 grade 3, stage T3 and operable stage T4 tumors should undergo an immediate inguinal node dissection because of the high probability of clinically occult lymph node invasion (in our material more than 50%). With respect to the extent of the node dissection, we found that the likelihood of spread to the contralateral and/or pelvic regions was related to the number of invaded nodes in the inguinal region. We recommend contralateral node dissection and unilateral pelvic node dissection when 2 or more positive nodes are found in the dissected groin specimen. Primary pelvic node dissection should be performed in patients who present initially with cytologically or biopsy proved positive inguinal nodes.(ABSTRACT TRUNCATED AT 400 WORDS)