Anatomic segmentation improves prostate cancer detection with artificial neural networks analysis of 1H magnetic resonance spectroscopic imaging

PURPOSE

To assess whether an artificial neural network (ANN) model is a useful tool for automatic detection of cancerous voxels in the prostate from (1)H-MRSI datasets and whether the addition of information about anatomical segmentation improves the detection of cancer.

MATERIALS AND METHODS

The Institutional Review Board approved this HIPAA-compliant study and waived informed consent. Eighteen men with prostate cancer (median age, 55 years; range, 36-71 years) who underwent endorectal MRI/MRSI before radical prostatectomy were included in this study. These patients had at least one cancer area on whole-mount histopathological map and at least one matching MRSI voxel suspicious for cancer detected. Two ANN models for automatic classification of MRSI voxels in the prostate were implemented and compared: model 1, which used only spectra as input, and model 2, which used the spectra plus information from anatomical segmentation. The models were trained, tested and validated using spectra from voxels that the spectroscopist had designated as cancer and that were verified on histopathological maps.

RESULTS

At ROC analysis, model 2 (AUC = 0.968) provided significantly better (P = 0.03) classification of cancerous voxels than did model 1 (AUC = 0.949).

CONCLUSION

Automatic analysis of prostate MRSI to detect cancer using ANN model is feasible. Application of anatomical segmentation from MRI as an additional input to ANN improves the accuracy of detecting cancerous voxels from MRSI.

Changes in pelvic organ prolapse surgery in the last decade among United States urologists

PURPOSE

Surgical correction of pelvic organ prolapse underwent transformation in the last decade. Training in pelvic organ prolapse surgery, the ease of mesh kit use, and Food and Drug Administration warnings about mesh have influenced practice patterns. We investigated trends in pelvic organ prolapse procedures.

MATERIALS AND METHODS

Case logs of pelvic organ prolapse procedures, mesh use and pessary placement were obtained from the American Board of Urology for 2003 to 2012. We evaluated associations between surgeon characteristics and the use of pelvic organ prolapse procedures.

RESULTS

Of 6,355 nonpediatric urologists applying for certification or recertification 2,192, representing a 10% annual sample of all urologists, reported performing pelvic organ prolapse procedures during the study period. The number of procedures increased steadily from 930 in 2003 to 6,978 in 2012. The number of colporrhaphies increased from 806 to 2,670 and the number of colpopexies increased from 32 to 1,414 between 2003 and 2012. The number of vaginal colpopexies increased from 24 to 1,016 during the study period. The number of sacrocolpopexies increased from 8 to 398 with exponential increases in laparoscopic sacrocolpopexy (282 cases by 2012). Mesh insertion increased from 10 cases reported by applicants in 2005 to 1,552 reported in 2012 (p <0.0005). Mesh revision, first reported in 2007 with 52 performed, consistently increased to 214 in 2012. Urologists trained in female urology performed a median of 16 pelvic organ prolapse procedures, double the number reported by surgeons trained in other urological fellowships. Urologists of the female gender also reported performing approximately 8 more procedures annually than male urologists.

CONCLUSIONS

The number of pelvic organ prolapse operations done by urologists increased dramatically in the last decade with a similar increase in mesh use. More colpopexies are now performed with laparoscopic sacrocolpopexy showing an exponential increase. The recent trend of mesh revision is notable with a much faster rate of increase than mesh insertion.

Midurethral sling is the dominant procedure for female stress urinary incontinence: analysis of case logs from certifying American Urologists

OBJECTIVE

To investigate contemporary trends in the use of midurethral sling procedures for the surgical correction of female stress urinary incontinence over the past decade.

METHODS

Annualized case log data for female incontinence surgeries from certifying and recertifying urologists were obtained from the American Board of Urology. Descriptive analysis of the number and type of cases per year was performed. Associations between surgeon characteristics and the use of female incontinence procedures were evaluated.

RESULTS

A total of 6355 nonpediatric urologists applied for certification or recertification between 2003 and 2012. Two-thirds (4185) reported performing any procedures for female incontinence. Procedures sharply increased from 4632 in 2003 to 7548 in 2004, then remained relatively stable between 2005 and 2012 (range, 8014-10,238 cases). Traditional procedures decreased from 17% of female incontinence procedures in 2003 to 5% in 2004 to <1% since 2010 (P <.0005). Midurethral sling procedures have risen sharply from 3210 procedures in 2003 to 7200 in 2012 (P <.0005). Endoscopic injection treatments have remained stable.

CONCLUSION

Midurethral slings have been widely adopted by urologists over the last decade. Increase in sling usage coincided with a drastic decline in traditional repairs, implying that the newer midurethral slings were replacing these traditional procedures for the treatment of female incontinence. In addition, the fact that the use of periurethral injections did not change significantly during this time period indicates that increased sling usage is responsible for most of the decline in traditional repairs.

MRI findings of radiation-induced changes in the urethra and periurethral tissues after treatment for prostate cancer

PURPOSE

To assess radiotherapy (RT)-induced changes in the urethra and periurethral tissues after treatment for prostate cancer (PCa).

METHODS AND MATERIALS

This retrospective study included 108 men (median age, 64 years; range, 43-87 years) who received external-beam radiotherapy (EBRT) and/or brachytherapy for PCa and underwent endorectal-coil MRI of the prostate within 180 days before RT and a median of 20 months (range, 2-62 months) after RT. On all MRIs, two readers independently measured the urethral length (UL) and graded the margin definition (MD) of the urethral wall and the signal intensities (SIs) of the urethral wall and pelvic muscles on 4-point scales.

RESULTS

The mean urethral length decreased significantly from pre- to post-RT MRI (from 15.2 to 12.6mm and from 14.4 to 12.9 mm for readers 1 and 2, respectively; both p-values <0.0001). Brachytherapy resulted in greater urethral shortening than EBRT. After RT, SI in the urethral wall increased in 57% (62/108) and 35% (38/108) of patients (readers 1 and 2, respectively). The frequency and magnitude of SI increase in pelvic muscles depended on muscle location. In the obturator internus muscle, SI increased more often after EBRT than after brachytherapy, while in the periurethral levator ani muscle SI increased more often after brachytherapy than after EBRT.

CONCLUSION

After RT for PCa, MRI shows urethral shortening and increased SI of the urethral wall and pelvic muscles in substantial percentages of patients.

Do margins matter? The influence of positive surgical margins on prostate cancer-specific mortality

BACKGROUND

Positive surgical margins (PSMs) in radical prostatectomy (RP) specimens are a frequent indication for adjuvant radiotherapy and are used as a measure of surgical quality. However, the association between PSMs and prostate cancer-specific mortality (CSM) is poorly defined.

OBJECTIVE

Analyze the association of PSMs with CSM, adjusting for fixed and time-dependent parameters.

DESIGN, SETTING, AND PARTICIPANTS

Fine and Gray competing risk regression analysis was used to model the clinical data and follow-up information of 11,521 patients treated by RP between 1987 and 2005. Two extended models were used that adjusted for the use of postoperative radiotherapy, which was handled as a time-dependent covariate. Postoperative radiotherapy was modeled as a single parameter and also as early and late therapy, based on the prostate-specific antigen level at the start of treatment (≤0.5 vs >0.5 ng/ml).

INTERVENTION

RP for clinically localized prostate cancer and selective use of secondary local and/or systemic therapy.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

The outcome measure was prostate cancer-specific mortality.

RESULTS AND LIMITATIONS

The 15-yr CSM rates for patients with PSMs and negative surgical margins were 10% and 6%, respectively (p<0.001). No significant association between PSM and CSM was observed in the conventional model with fixed covariates (hazard ratio [HR]: 1.04; 95% confidence interval [CI], 0.7-1.5; p=0.8) or in the two extended models that adjusted for postoperative radiotherapy (HR: 0.96; 95% CI, 0.7-1.4; p=0.9), or early and late postoperative radiotherapy (HR: 1.01; 95% CI, 0.7-1.4; p=0.9).

CONCLUSIONS

PSMs alone are not associated with a significantly increased risk of CSM within 15 yr of RP. However, urologists should continue to strive to avoid PSMs, as they increase a man's risk of biochemical recurrence and need for secondary therapy and may be a source of considerable patient anxiety.

Positive surgical margins after radical prostatectomy: a systematic review and contemporary update

CONTEXT

The clinical significance of positive surgical margins (PSMs) in radical prostatectomy (RP) specimens and the management of affected patients remain unclear.

OBJECTIVE

To address pitfalls in the pathologic interpretation of margin status; provide an update on the incidence, predictors, and long-term oncologic implications of PSMs in the era of robot-assisted laparoscopic RP (RALRP); and suggest a practical evidence-based approach to patient management.

EVIDENCE ACQUISITION

A systematic review of the literature was performed in April 2013 using Medline/PubMed, Web of Science, and Scopus databases and the Cochrane Database of Systematic Reviews. Studies focusing on PSMs in RP pertinent to the objectives of this review were included. Particular attention was paid to publications within the last 5 yr and those concerning RALRP.

EVIDENCE SYNTHESIS

A total of 74 publications were retrieved. Standardized measures to overcome variability in the pathologic interpretation of surgical margins have recently been established by the International Society of Urological Pathology. The average rate of PSMs in contemporary RALRP series is 15% (range: 6.5-32%), which is higher in men with a more advanced pathologic stage and equivalent to the rate reported in prior open and laparoscopic prostatectomy series. The likelihood of PSMs is strongly influenced by the surgeon's experience irrespective of the surgical approach. Technical modifications using the robotic platform and the role of frozen-section analysis to reduce the margin positivity rate continue to evolve. Positive margins are associated with a twofold increased hazard of biochemical relapse, but their association with more robust clinical end points is controversial. Level 1 evidence suggests that adjuvant radiation therapy (RT) may favorably affect prostate-specific antigen recurrence rates, but whether the therapy also affects systemic progression, prostate cancer-specific mortality, and overall survival remains debatable.

CONCLUSIONS

Although positive margins in prostate cancer are considered an adverse oncologic outcome, their long-term impact on survival is highly variable and largely influenced by other risk modifiers. Adjuvant RT appears to be effective, but further study is required to determine whether early salvage RT is an equivalent alternative.

Risk of incisional hernia after minimally invasive and open radical prostatectomy

PURPOSE

The number of radical prostatectomies has increased. Many urologists have shifted from the open surgical approach to minimally invasive techniques. It is not clear whether the risk of post-prostatectomy incisional hernia varies by surgical approach.

MATERIALS AND METHODS

In the linked Surveillance, Epidemiology and End Results (SEER)-Medicare data set we identified men 66 years old or older who were treated with minimally invasive or open radical prostatectomy for prostate cancer diagnosed from 2003 to 2007. The main study outcome was incisional hernia repair, as identified in Medicare claims after prostatectomy. We also examined the frequency of umbilical, inguinal and other hernia repairs.

RESULTS

We identified 3,199 and 6,795 patients who underwent minimally invasive and open radical prostatectomy, respectively. The frequency of incisional hernia repair was 5.3% at a median 3.1-year followup in the minimally invasive group and 1.9% at a 4.4-year median followup in the open group, corresponding to an incidence rate of 16.1 and 4.5/1,000 person-years, respectively. Compared to the open technique, the minimally invasive procedure was associated with more than a threefold increased risk of incisional hernia repair when controlling for patient and disease characteristics (adjusted HR 3.39, 95% CI 2.63-4.38, p<0.0001). Minimally invasive radical prostatectomy was associated with an attenuated but increased risk of any hernia repair compared with open radical prostatectomy (adjusted HR 1.48, 95% CI 1.29-1.70, p<0.0001).

CONCLUSIONS

Minimally invasive radical prostatectomy was associated with a significantly increased risk of incisional hernia compared with open radical prostatectomy. This is a potentially remediable complication of prostate cancer surgery that warrants increased vigilance with respect to surgical technique.

Long-term outcomes of patients with lymph node metastasis treated with radical prostatectomy without adjuvant androgen-deprivation therapy

BACKGROUND

The presence of lymph node metastasis (LNM) at radical prostatectomy (RP) is associated with poor outcome, and optimal treatment remains undefined. An understanding of the natural history of node-positive prostate cancer (PCa) and identifying prognostic factors is needed.

OBJECTIVE

To assess outcomes for patients with LNM treated with RP and lymph node dissection (LND) alone.

DESIGN, SETTING, AND PARTICIPANTS

We analyzed data from a consecutive cohort of 369 men with LNM treated at a single institution from 1988 to 2010.

INTERVENTION

RP and extended LND.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Our primary aim was to model overall survival, PCa-specific survival, metastasis-free progression, and freedom from biochemical recurrence (BCR). We used univariate Cox proportional hazard regression models for survival outcomes. Multivariable Cox proportional hazard regression models were used for freedom from metastasis and freedom from BCR, with prostate-specific antigen, Gleason score, extraprostatic extension, seminal vesical invasion, surgical margin status, and number of positive nodes as predictors.

RESULTS AND LIMITATIONS

Sixty-four patients with LNM died, 37 from disease. Seventy patients developed metastasis, and 201 experienced BCR. The predicted 10-yr overall survival and cancer-specific survival were 60% (95% confidence interval [CI], 49-69) and 72% (95% CI, 61-80), respectively. The 10-yr probability of freedom from distant metastasis and freedom from BCR were 65% (95% CI, 56-73) and 28% (95% CI, 21-36), respectively. Higher pathologic Gleason score (>7 compared with ≤ 7; hazard ratio [HR]: 2.23; 95% CI, 1.64-3.04; p < 0.0001) and three or more positive lymph nodes (HR: 2.61; 95% CI, 1.81-3.76; p < 0.0001) were significantly associated with increased risk of BCR on multivariable analysis. The retrospective nature and single-center source of data are study limitations.

CONCLUSIONS

A considerable subset of men with LNM remained free of disease 10 yr after RP and extended LND alone. Patients with pathologic Gleason score <8 and low nodal metastatic burden represent a favorable group. Our data confirm prior findings and support a plea for risk subclassification for patients with LNM.

Pelvic lymph node dissection during robot-assisted radical prostatectomy: efficacy, limitations, and complications-a systematic review of the literature

CONTEXT

Pelvic lymph node dissection (PLND) in prostate cancer is the most effective method for detecting lymph node metastases. However, a decline in the rate of PLND during radical prostatectomy (RP) has been noted. This is likely the result of prostate cancer stage migration in the prostate-specific antigen-screening era, and the introduction of minimally invasive approaches such as robot-assisted radical prostatectomy (RARP).

OBJECTIVE

To assess the efficacy, limitations, and complications of PLND during RARP.

EVIDENCE ACQUISITION

A review of the literature was performed using the Medline, Scopus, and Web of Science databases with no restriction of language from January 1990 to December 2012. The literature search used the following terms: prostate cancer, radical prostatectomy, robot-assisted, and lymph node dissection.

EVIDENCE SYNTHESIS

The median value of nodal yield at PLND during RARP ranged from 3 to 24 nodes. As seen in open and laparoscopic RP series, the lymph node positivity rate increased with the extent of dissection during RARP. Overall, PLND-only related complications are rare. The most frequent complication after PLND is symptomatic pelvic lymphocele, with occurrence ranging from 0% to 8% of cases. The rate of PLND-associated grade 3-4 complications ranged from 0% to 5%. PLND is associated with increased operative time. Available data suggest equivalence of PLND between RARP and other surgical approaches in terms of nodal yield, node positivity, and intraoperative and postoperative complications.

CONCLUSIONS

PLND during RARP can be performed effectively and safely. The overall number of nodes removed, the likelihood of node positivity, and the types and rates of complications of PLND are similar to pure laparoscopic and open retropubic procedures.

Back to baseline: erectile function recovery after radical prostatectomy from the patients' perspective

INTRODUCTION

A variety of erectile function recovery (EFR) rates are reported post-radical prostatectomy (RP), with some suggesting EFR rates over 90% [1]. Clinical experience suggests that patients view EFR as getting back to their baseline (BTB) erectile functioning (EF) without the use of medication.

AIM

This study explores EFR defined as BTB.

METHOD

Men pre-RP and 24 months post-RP completed the Erectile Function Domain (EFD) of the International Index of Erectile Function and one question on phosphodiesterase type 5 inhibitor (PDE5i) use. Men using a PDE5i at baseline were excluded.

MAIN OUTCOME MEASURES

At 24 m, "back to baseline" was defined as achieving the baseline EFD score (within 1 point or higher). Analyses included descriptive statistics, chi-square, and logistic regression.

RESULTS

One hundred eighty men had an average age at RP of 59 (SD = 7) years. When including men who were using a PDE5i at 24 months, 43% (N = 78, 95% CI: 36-51%) returned BTB. When considering BTB without the use of a PDE5i, 22% (N = 39, 95% CI: 16% to 28%) returned BTB. When focusing on a subset of men with baseline EFD ≥ 24 (N = 132), 36% (N = 47, 95% CI: 28% to 44%) returned BTB at 24 months using a PDE5i and 16% (N = 21, 95% CI: 11% to 23%) without the use of a PDE5i. For this group, there was a significant difference by age (<60 years, 23% vs. ≥ 60 years, 4%, P < 0.001), which remained a significant predictor (OR = 6.25, 95% CI: 1.88 to 50, P < 0.001) in multivariable analysis.

CONCLUSIONS

Twenty-two percent of the entire sample and 16% of the men with functional (EFD ≥ 24) baseline erections returned to BTB EF without the use of medication. Only 4% of men who were ≥ 60 years old with functional erections pre-surgery achieved BTB EF. Although gaining partial EF is also important, men pre-RP should be educated on EFR and the chance of "back to baseline" EF.

A case-mix-adjusted comparison of early oncological outcomes of open and robotic prostatectomy performed by experienced high volume surgeons

OBJECTIVE

To compare early oncological outcomes of robot assisted laparoscopic prostatectomy (RALP) and open radical prostatectomy (ORP) performed by high volume surgeons in a contemporary cohort.

METHODS

We reviewed patients who underwent radical prostatectomy for prostate cancer by high volume surgeons performing RALP or ORP. Biochemical recurrence (BCR) was defined as PSA ≥ 0.1 ng/mL or PSA ≥ 0.05 ng/mL with receipt of additional therapy. A Cox regression model was used to evaluate the association between surgical approach and BCR using a predictive model (nomogram) based on preoperative stage, grade, volume of disease and PSA. To explore the impact of differences between surgeons, multivariable analyses were repeated using surgeon in place of approach.

RESULTS

Of 1454 patients included, 961 (66%) underwent ORP and 493 (34%) RALP and there were no important differences in cancer characteristics by group. Overall, 68% of patients met National Comprehensive Cancer Network (NCCN) criteria for intermediate or high risk disease and 9% had lymph node involvement. Positive margin rates were 15% for both open and robotic groups. In a multivariate model adjusting for preoperative risk there was no significant difference in BCR rates for RALP compared with ORP (hazard ratio 0.88; 95% CI 0.56-1.39; P = 0.6). The interaction term between nomogram risk and procedure type was not statistically significant. Using NCCN risk group as the covariate in a Cox model gave similar results (hazard ratio 0.74; 95% CI 0.47-1.17; P = 0.2). The interaction term between NCCN risk and procedure type was also non-significant. Differences in BCR rates between techniques (4.1% vs 3.3% adjusted risk at 2 years) were smaller than those between surgeons (2.5% to 4.8% adjusted risk at 2 years).

CONCLUSIONS

In this relatively high risk cohort of patients undergoing radical prostatectomy we found no evidence to suggest that ORP resulted in better early oncological outcomes then RALP. Oncological outcome after radical prostatectomy may be driven more by surgeon factors than surgical approach.

Can transrectal needle biopsy be optimised to detect nearly all prostate cancer with a volume of ≥0.5 mL? A three-dimensional analysis

OBJECTIVE

To investigate whether transrectal needle biopsy can be optimised to detect nearly all prostate cancer with a tumour volume (TV) of ≥0.5 mL.

MATERIALS AND METHODS

Retrospectively analysed 109 whole-mounted and entirely submitted radical prostatectomy specimens with prostate cancer. All tumours in each prostate were outlined on whole-mount slides and digitally scanned to produce tumour maps. Tumour map images were exported to three-dimensional (3D) slicer software (http://www.slicer.org) to develop a 3D-prostate cancer model. In all, 20 transrectal biopsy schemes involving two to 40 cores and two to six anteriorly directed biopsy (ADBx) cores (including transition zone, TZ) were simulated, as well as models with various biopsy cutting lengths. Detection rates for tumours of different volumes were determined for the various biopsy simulation schemes.

RESULTS

In 109 prostates, 800 tumours were detected, 90 with a TV of ≥0.5 mL (mean TV 0.24 mL). Detection rate for tumours with a TV of ≥0.5 mL plateaued at 77% (69/90) using a 12-core (3 × 4) scheme, standard 17-mm biopsy cutting length without ADBx cores. In all, 20 of 21 (95%) tumours with a TV of ≥0.5 mL not detected by this scheme originated in the anterior peripheral zone or TZ. Increasing the biopsy cutting length and depth/number of ADBx cores improved the detection rate for tumours with a TVof ≥0.5 mL in the 12-core scheme. Using a 22-mm cutting length and a 12-core scheme with additional volume-adjusted ADBx cores, 100% of ≥0.5 mL tumours in prostates ≤ 50 mL in volume and 94.7% of ≥0.5 mL tumours in prostates > 50 mL in volume were detected.

CONCLUSIONS

Our 3D-prostate cancer model analysis suggests that nearly all prostate cancers with a TV of ≥0.5 mL can be detected by 14-18 transrectal needle-biopsy cores. Using longer biopsy cutting lengths and increasing the depth and number of ADBx cores (including TZ) according to prostate volume are necessary as well.

Performance characteristics of MR imaging in the evaluation of clinically low-risk prostate cancer: a prospective study

PURPOSE

To prospectively evaluate diagnostic performance of T2-weighted magnetic resonance (MR) imaging and MR spectroscopic imaging in detecting lesions stratified by pathologic volume and Gleason score in men with clinically determined low-risk prostate cancer.

MATERIALS AND METHODS

The institutional review board approved this prospective, HIPAA-compliant study. Written informed consent was obtained from 183 men with clinically low-risk prostate cancer (cT1-cT2a, Gleason score≤6 at biopsy, prostate-specific antigen [PSA] level<10 ng/mL [10 μg/L]) undergoing MR imaging before prostatectomy. By using a scale of 1-5 (score 1, definitely no tumor; score 5, definitely tumor), two radiologists independently scored likelihood of tumor per sextant on T2-weighted images. Two spectroscopists jointly recorded locations of lesions with metabolic features consistent with tumor on MR spectroscopic images. Whole-mount step-section histopathologic analysis constituted the reference standard. Diagnostic performance at sextant level (T2-weighted imaging) and detection sensitivities (T2-weighted imaging and MR spectroscopic imaging) for lesions of 0.5 cm3 or larger were calculated.

RESULTS

For T2-weighted imaging, areas under the receiver operating characteristic curves for sextant-level detection were 0.77 (reader 1) and 0.82 (reader 2). For lesions of ≥0.5 cm3 and, 1

CONCLUSION

In men with clinically low-risk prostate cancer, detection of lesions of <1 cm3 with T2-weighted imaging is significantly dependent on lesion Gleason score; detection of lesions of ≥1 cm3 is significantly better than detection of smaller lesions and is not affected by lesion Gleason score. The role of MR spectroscopic imaging alone in this population is limited.

Collapse

Key Words Collapse

MESH Headings

• Adult
• Aged
• Humans
• Image Enhancement/methods
• Image Interpretation, Computer-Assisted/methods
• Male
• Middle Aged
• Preoperative Care/methods
• Prospective Studies
• Prostatic Neoplasms/pathology
• Reproducibility of Results
• Risk Factors
• Sensitivity and Specificity

Collapse

Grants

• R01 CA076423 NCI NIH HHS

Collapse