An MRI-Based Convolutional Neural Network to Predict Biochemical Recurrence Following Radiotherapy for Intermediate and High-Risk Localized Prostate Cancer

PURPOSE/OBJECTIVE(S)

The risk of biochemical recurrence (BCR) following radiotherapy (RT) for localized prostate cancer (LPCa) varies considerably within risk stratification groups defined by classic clinical and pathologic variables; there is an unmet need for low-cost tools that more robustly predict BCR and allow for individualized therapy. Published imaging-based algorithms for BCR prediction after RT are limited to hand-crafted radiomics and/or small cohorts. We aimed to develop a deep learning model to predict BCR at 5 years after RT for intermediate and high risk LPCa using pre-treatment T2-weighted (T2W) MRI.

MATERIALS/METHODS

Patients with intermediate and high risk LPCa treated with radical RT at our institution between 2010 and 2015 were included. We excluded those who did not have a pre-treatment T2W-MRI and those with less than 5 years of follow-up. The Phoenix definition for BCR was used. The dataset (DS1) was split into training (70%), validation (20%), and test (10%) sets using a stratified technique. A U-Net model for prostate segmentation was trained and tested on a separate annotated prostate T2W-MRI dataset (DS2) of 225 patients from our institution. The U-Net model was then used to segment the whole prostate gland on the MRI images of DS1, and the segmented images were fed into four 2D convolutional neural networks (CNNs) using different network architectures and regularization techniques (VGG blocks with batch normalization, dropout, and max pooling layers) to predict BCR at 5 years. The CNNs were evaluated using the area under the receiver operating characteristic curve (AUC) on the test set. For benchmarking, three machine learning classifiers (Random Forest, Logistic Regression, and Support Vector Machines) were developed using the 5 most important features selected by Mean Decrease in Impurity from a set of 18 clinical variables.

RESULTS

A total of 189 patients were included in DS1. Androgen deprivation therapy (ADT) was received by 83.6% of patients. BCR was identified in 26% of the cases. The Dice score for the U-Net segmentation model was 78% on the test set of DS2. The AUC achieved by the different CNNs for predicting BCR ranged between 0.53 and 0.75. The best performing CNN consisted of 3 convolutional layers, the first two followed by max-pooling layers, a flattening layer, a dense layer, and an output layer with softmax activation function. The best clinical model was a Random Forest algorithm with an AUC of 0.70. The selected clinical variables by decreasing feature importance were: age, time to nadir PSA, pre-treatment PSA, percentage of positive biopsy cores at diagnosis, and nadir PSA.

CONCLUSION

We developed a deep learning model based on pre-treatment T2W-MRI to predict BCR at 5 years following radical RT for intermediate and high-risk LPCa. This CNN outperformed a model based on clinical variables and warrants further validation in external cohorts.

5-Year Renal Function Outcomes after SABR for Primary Renal Cell Carcinoma: A Report from the International Radiosurgery Oncology Consortium of the Kidney (IROCK)

PURPOSE/OBJECTIVE(S)

Renal cell carcinoma (RCC) presents uncommonly in patients with a congenital solitary kidney or prior contralateral nephrectomy. The objective of this study was to compare renal function outcomes of stereotactic ablative body radiotherapy (SABR) in patients with solitary vs. bilateral kidneys.

MATERIALS/METHODS

Patients with primary RCC with ≥2 years of follow-up at 12 participating International Radiosurgery Consortium for Kidney (IROCK) institutions were included. Patients with upper tract urothelial carcinoma or metastatic disease were excluded. Renal function was measured by estimated glomerular filtration rate (eGFR). For patients where eGFR was not recorded, Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation was used to estimate eGFR based on known creatinine. Baseline characteristics and renal function outcomes were compared between solitary vs. bilateral kidneys. Multivariable logistic regression was used to identify factors predictive of eGFR decline ≥ 15 mL/min and any eGFR increase evaluated at 1-year post-SABR.

RESULTS

One hundred and ninety patients with solitary (n = 56) or bilateral kidneys (n = 134) underwent SABR and were followed for a median of 5.0 years (IQR: 3.4-6.8). Pre-SABR eGFR (mean ± SD) was similar in patients with solitary (61.1 ± 23.2 mL/min) vs. bilateral kidneys (58.0 ± 22.3 mL/min, p = 0.324). Mean tumor size was 3.70 ± 1.40 cm in solitary and 4.35 ± 2.50 cm in bilateral kidneys (p = 0.026). After SABR, an initial compensatory increase in eGFR was observed in both cohorts (22.7% solitary and 17.7% bilateral at 1 year). This compensatory increase persisted in patients with bilateral but not a solitary kidney (10.3% vs. 0% at 3-years and 21.1% vs. 0% at 5-years, respectively). At 5-years post-SABR, eGFR decreased by -14.5 ± 7.6 in solitary and -13.3 ± 15.9 mL/min in bilateral kidneys (p = 0.665). At all timepoints assessed, there were no significant differences in eGFR decline between solitary vs. bilateral cohorts (all p > 0.05). There were also no significant differences in post-SABR end-stage renal disease (7.1% vs. 6.7%) or dialysis (3.6% vs. 3.7%) in solitary vs. bilateral, respectively. Multivariable analysis demonstrated that increasing tumor size (OR per 1 cm: 1.57; 95% CI: 1.14-2.16, p = 0.006) and baseline eGFR (OR per 10 mL/min: 1.30; 95% CI: 1.02-1.66, p = 0.034) was more likely to be associated with eGFR decline ≥ 15 mL/min. There was no significant association between solitary vs. bilateral kidney and eGFR decline (OR: 1.22; 95% CI: 0.45-3.34, p = 0.693).

CONCLUSION

There was no observed difference between renal function outcomes in patients with a solitary vs. bilateral kidneys. While larger tumor size may increase the risk of eGFR decline post-SABR, treatment of a solitary kidney does not appear to increase the risk of renal dysfunction long-term.

Impact of Prior Radiation Therapy on Bone Mineral Density Change Over Time: Secondary Analysis of the Control Arm of a Phase III Randomized Trial

PURPOSE/OBJECTIVE(S)

Retrospective studies have demonstrated that pelvic radiation therapy (RT) can lead to decreased bone mineral density (BMD) and increased risk of fracture. This is more relevant for men with prostate cancer (PCa) who often receive androgen deprivation therapy (ADT) in conjunction with RT. We performed a post-hoc secondary analysis of publicly available data of the control arm of a phase III randomized controlled study (NCT00089674) to determine if history of prior pelvic RT affects change in BMD over time in non-metastatic PCa patients treated with ADT.

MATERIALS/METHODS

In this study, PCa patients with age ≥70 years or <70 years with low BMD (T-score <-1) or history of osteoporotic fracture, on ADT for at least 12 months were randomized to receive densoumab vs. placebo every 6 months for 3 years. Additionally, all patients received daily vitamin D and calcium supplementation. Randomization was stratified by duration of prior ADT (≤6 months vs >6 months) and age (<70 vs ≥70 years). BMD was measured at baseline, and at months 1, 3, 6, 12, 24, and 36 with blind reading by central reviewer. To model the effect of prior pelvic RT on dynamic change in BMD in the hip, lumbar spine, and femoral neck, we applied separate multivariate linear mixed effect models for each site. Age, ECOG performance score, history and number of prior fractures, smoking history, and years from initial cancer diagnosis were included as fixed covariates while patients were included as random intercepts.

RESULTS

Among 734 patients who were randomized to the control arm, 563 participants with baseline and at least one post baseline assessment of BMD were eligible for this analysis. Overall, 34.4% (n = 194) received prior RT. We did not find any significant association of dynamic change in BMD with receipt of prior pelvic RT for left femoral neck (p = 0.7), total hip (p = 0.8), and lumbar spine (p = 0.5), respectively. At 36 months, there was no significant association of prior RT with percent change in BMD in femoral neck (odds ratio [OR]: 0.85; 95% confidence interval [CI]: 0.30-2.41), total hip (OR: 0.96; 95% CI: 0.43-2.15), and lumbar spine (OR: 2.01; 95% CI: 0.63-6.45). However, note should be made of the opposite direction of association of prior RT with percent BMD change at 36 months for femoral neck and hip versus lumbar spine.

CONCLUSION

In this exploratory analysis of the control arm of a phase III randomized trial, we did not find sufficient evidence of an association between prior pelvic RT and dynamic changes in BMD in femoral neck, hip, and lumbar spine over time in men with non-metastatic PCa and low BMD at baseline. This analysis should be interpreted cautiously considering its post-hoc nature with likely inadequate power, the possibility of selection bias, lack of information on receipt of prior ADT, and missing data in longitudinal assessments.

Association of Prostate Specific Antigen Kinetics after Testosterone Recovery with Subsequent Recurrence: Secondary Analysis of a Phase III Randomized Controlled Trial

PURPOSE/OBJECTIVE(S)

The combination of short-term androgen deprivation therapy (ST-ADT) with prostate radiotherapy (RT) is a standard of care for patients with localized prostate cancer (LPCa). After cessation of ST-ADT, it takes about 8 to 10 months for the testosterone (T) to recover to supracastrate levels, which could drive changes in PSA kinetics. It largely remains unknown whether early changes in PSA kinetics after T recovery could predict for subsequent biochemical relapse.

MATERIALS/METHODS

We performed a secondary analysis of a phase III randomized controlled trial in which patients with newly diagnosed LPCa with Gleason score £7, clinical stage T1b to T3a, and PSA <30 ng/mL were randomly allocated to neoadjuvant and concurrent ADT for 6 months starting 4 months before prostate RT (76 Gy in 38 fractions over 7.5 weeks) or concurrent and adjuvant ADT for 6 months starting simultaneously with prostate RT. Clinical assessment and laboratory investigations were repeated 1 month after completion of ADT, every 4 months for the first 2 years, every 6 months for the next 3 years, and annually thereafter. We calculated the PSA doubling time (PSADT) based on PSA values up to 18 months after recovery of T to a supracastrate level (>50 ng/dL). Patients with ³3 PSA measurements after T recovery to supracastrate level were included in this analysis. Fine and Gray cumulative incidence of biochemical recurrence (BCR) was calculated in patients with PSADT at or above median versus below median. Deaths were considered as competing events. All endpoints were calculated from the time of T recovery to supracastrate level. Subdistribution hazard ratios (sHR) with 95% confidence intervals (CI) were estimated for association of PSADT with relative incidence of recurrence using competing risk regression after adjusting for tumor stage, pre-treatment PSA, Gleason score, treatment regimen, and age at randomization.

RESULTS

Overall, 311 patients were eligible for this analysis. Median PSADT was 8 months. Cumulative incidence of BCR at 10 years was 31.0% and 20.7% in patients with PSADT <8 months and ³8 months, respectively. Longer PSADT was associated with a significantly lower risk of cumulative incidence of BCR (sHR for PSADT as a continuous variable 0.43, 95% CI: 0.28-0.66; sHR for PSADT ³8 months 0.54, 95% CI: 0.30-0.99). After adjustment for time to recovery of T to supracastrate level in addition to the aforementioned variables, longer PSADT (³8 months) was associated with lower risk of cumulative incidence of BCR (sHR: 0.53, 95% CI: 0.27-1.01).

CONCLUSION

These findings suggest that early PSA kinetics within 18 months of recovery of T to a supracastrate level predict for subsequent biochemical failure. Taking account of early changes in PSA after testosterone recovery may allow for recognition of potential failures earlier in the disease course and thereby permit greater personalization of management decisions.

Impact of Concomitant Medications on Biochemical Outcome in Localised Prostate Cancer Treated with Radiotherapy and Androgen Deprivation Therapy

AIMS

Several classes of concomitant medications have been shown to affect oncological outcomes in patients with prostate cancer (PCa). We assessed the association between the use of commonly prescribed concomitant medications and biochemical relapse-free survival (bRFS) in patients with localised PCa treated with radiotherapy and androgen deprivation therapy (ADT).

MATERIALS AND METHODS

A secondary pooled analysis of two phase III randomised trials was carried out. In the first trial, patients with localised PCa with clinical stage T1b-T3, prostate-specific antigen <30 ng/ml and Gleason score ≤7 were treated with radical radiotherapy and 6 months of ADT starting 4 months before or concomitantly with radiotherapy. In the second trial, patients with high-risk PCa were treated with radical radiotherapy and 36 months of ADT with randomisation to three-dimensional conformal or intensity-modulated radiotherapy. Information on concomitant medications was collected from the medical record. Univariable and multivariable Cox regression was used to identify factors associated with bRFS.

RESULTS

Overall, 486 patients were evaluable. The median follow-up was 125 months; 10-year bRFS was 83.7%. On univariable analysis, receipt of metformin was significantly associated with worse bRFS. Ten-year bRFS was 73% and 85% for patients with and without concomitant metformin (adjusted hazard ratio 2.11, 95% confidence interval 1.03-4.33). Similar evidence of an association was observed with sulfonamide-based α1-receptor blockers (adjusted hazard ratio 2.72, 95% confidence interval 1.31-5.66). However, no such association was seen with receipt of quinazoline-based α1-receptor blockers (adjusted hazard ratio 1.09, 95% confidence interval 0.42-2.82). There was no significant association between bRFS and receipt of all other medication classes considered.

CONCLUSIONS

In this population of patients with localised PCa treated with radiotherapy and ADT, receipt of concomitant metformin and sulfonamide-based α1-receptor blockers was associated with inferior biochemical outcome. Randomised trials are required to assess the true effect of these medications on oncological outcomes in localised PCa.

Who Dies From Prostate Cancer? An Analysis of the Surveillance, Epidemiology and End Results Database

AIMS

To characterise the presenting features of those who ultimately die from prostate cancer (PCa).

MATERIALS AND METHODS

The study population consisted of patients in the Surveillance, Epidemiology and End Results (SEER) Program database diagnosed with PCa between 1990 and 2015. Patients were assigned to the following clinical risk groups: low-risk localised (LRL), intermediate-risk localised (IRL), high-risk localised (HRL), node-positive and metastatic (M1). Before 2004, in the absence of prostate-specific antigen (PSA) and Gleason score data, patients with cT1-T2aN0M0 and low-grade PCa were classified as LRL, those with cT3-4N0M0 or high-grade PCa were classified as HRL and all others with N0M0 disease were classified as IRL. The primary aim was to describe the risk group distribution of those who ultimately died from PCa compared with those who were diagnosed with PCa over the study period. A secondary aim was to estimate PCa-specific survival (PCSS) and evaluate the association of risk group with PCSS.

RESULTS

Among a total of 811 487 patients who were diagnosed with PCa, data sufficient for risk group determination were present in 635 733 patients. The median follow-up was 83 months. The overall risk group distribution at diagnosis was as follows: LRL 10.5%, IRL 49.7%, HRL 34.8%, node-positive 1.5% and M1 3.5%. The risk group distribution of those who died from PCa was 3.9%, 29.4%, 40.9%, 3.2% and 22.8%, respectively. Compared with LRL PCa, the adjusted hazard ratio (95% confidence interval) for PCSS was 1.40 (1.33-1.46) in IRL, 3.76 (3.60-3.93) in HRL, 11.87 (11.14-12.65) in node-positive and 37.12 (35.43-38.88) in M1.

CONCLUSIONS

In this large contemporary cohort, patients with M1, node-positive and HRL disease accounted for two-thirds of all deaths from PCa. De novo metastatic PCa was associated with an approximately 40-fold increased risk of death from PCa compared with LRL PCa. Efforts to improve PCSS will therefore depend largely on improvements in therapy in those with M1, node-positive and HRL disease.

Bone Health and Bone-targeted Therapies for Prostate Cancer: a Programme in Evidence-based Care - Cancer Care Ontario Clinical Practice Guideline

AIMS

To make recommendations with respect to bone health and bone-targeted therapies in men with prostate cancer.

MATERIALS AND METHODS

A systematic review was carried out by searching MEDLINE, EMBASE and the Cochrane Library from inception to January 2016. Systematic reviews and randomised-controlled trials were considered for inclusion if they involved therapies directed at improving bone health or outcomes such as skeletal-related events, pain and quality of life in patients with prostate cancer either with or without metastases to bone. Therapies included medications, supplements or lifestyle modifications alone or in combination and were compared with placebo, no treatment or other agents. Disease-targeted agents such as androgen receptor-targeted and chemotherapeutic agents were excluded. Recommendations were reviewed by internal and external review groups.

RESULTS

In men with prostate cancer receiving androgen deprivation therapy, baseline bone mineral density testing is encouraged. Denosumab should be considered for reducing the risk of fracture in men on androgen deprivation therapy with an increased fracture risk. Bisphosphonates were effective in improving bone mineral density, but the effect on fracture was inconclusive. No medication is recommended to prevent the development of first bone metastasis. Denosumab and zoledronic acid are recommended for preventing or delaying skeletal-related events in men with metastatic castration-resistant prostate cancer. Radium-223 is recommended for reducing symptomatic skeletal events and prolonging survival in men with symptomatic metastatic castration-resistant prostate cancer.

CONCLUSIONS

The recommendations represent a current standard of care that is feasible to implement, with outcomes valued by clinicians and patients.

Effect on therapeutic ratio of planning a boosted radiotherapy dose to the dominant intraprostatic tumour lesion within the prostate based on multifunctional MR parameters

OBJECTIVE

To demonstrate the feasibility of an 8-Gy focal radiation boost to a dominant intraprostatic lesion (DIL), identified using multiparametric MRI (mpMRI), and to assess the potential outcome compared with a uniform 74-Gy prostate dose.

METHODS

The DIL location was predicted in 23 patients using a histopathologically verified model combining diffusion-weighted imaging, dynamic contrast-enhanced imaging, T2 maps and three-dimensional MR spectroscopic imaging. The DIL defined prior to neoadjuvant hormone downregulation was firstly registered to MRI-acquired post-hormone therapy and subsequently to CT radiotherapy scans. Intensity-modulated radiotherapy (IMRT) treatment was planned for an 8-Gy focal boost with 74-Gy dose to the remaining prostate. Areas under the dose-volume histograms (DVHs) for prostate, bladder and rectum, the tumour control probability (TCP) and normal tissue complication probabilities (NTCPs) were compared with those of the uniform 74-Gy IMRT plan.

RESULTS

Deliverable IMRT plans were feasible for all patients with identifiable DILs (20/23). Areas under the DVHs were increased for the prostate (75.1 ± 0.6 vs 72.7 ± 0.3 Gy; p < 0.001) and decreased for the rectum (38.2 ± 2.5 vs 43.5 ± 2.5 Gy; p < 0.001) and the bladder (29.1 ± 9.0 vs 36.9 ± 9.3 Gy; p < 0.001) for the boosted plan. The prostate TCP was increased (80.1 ± 1.3 vs 75.3 ± 0.9 Gy; p < 0.001) and rectal NTCP lowered (3.84 ± 3.65 vs 9.70 ± 5.68 Gy; p = 0.04) in the boosted plan. The bladder NTCP was negligible for both plans.

CONCLUSION

Delivery of a focal boost to an mpMRI-defined DIL is feasible, and significant increases in TCP and therapeutic ratio were found.

ADVANCES IN KNOWLEDGE

The delivery of a focal boost to an mpMRI-defined DIL demonstrates statistically significant increases in TCP and therapeutic ratio.

Abstract P4-16-06: Radiotherapy to the Primary Tumor Is Associated with Improved Survival in Stage IV Breast Cancer

Background: In patients found to have metastatic disease at the time of breast cancer diagnosis, the role of local therapy is undefined. Numerous retrospective analyses have suggested that surgery and/or external beam radiotherapy (EBRT) directed at the primary tumor may improve overall survival (OS). All these analyses, however, are subject to significant selection bias. The current retrospective analysis of a large registry dataset attempts to limit the effect of this bias.

Methods: The study population consisted of women in the Surveillance, Epidemiology, and End Results (SEER) program database diagnosed with stage IV breast cancer between 1988 and 2009. Only those patients for whom surgery to the primary tumor was recommended but was not undertaken (due to patient refusal or other uncategorized reasons) were included. In this population of patients deemed candidates for surgery, the association between receipt of primary tumor-directed EBRT and overall survival was studied. Descriptive statistics were used to characterize the study population. OS was estimated using the Kaplan-Meier (KM) method. Univariate and multivariate Cox regression were used to identify factors associated with OS.

Results: A total of 3,529 cases were analyzed. EBRT was received in 768 cases. Median age at diagnosis was 68 years (IQR, 56–79 years). Median follow-up by reverse KM estimate was 98 months (range, 0–252 months). On univariate analysis, EBRT was associated with improved OS (hazard ratio 0.80, 95% CI 0.74–0.87, p &lt; 0.001). 1-year, 3-year, and 5-year OS was 56.9%, 24.2%, and 10.7% respectively in those receiving EBRT and 44.3%, 16.6%, and 7.2% respectively in those not receiving EBRT. Median OS in those receiving EBRT was 15 months compared to 7 months in those not receiving EBRT. In a multivariate Cox model taking into account receipt of EBRT, age at diagnosis, year of diagnosis, ethnicity, number of primary cancers, estrogen and progesterone receptor status, histologic grade, and size of primary tumor, EBRT remained significantly associated with improved survival (hazard ratio 0.86, 95% CI 0.76–0.97, p = 0.011).

Conclusions: In a population of women presenting with metastatic breast cancer, all of whom were deemed candidates for surgery to the primary tumor but who did not undergo surgery, receipt of EBRT was associated with improved OS. The observed 8-month absolute difference in median OS is clinically significant. This analysis could not account for performance status, extent of metastatic disease, co-morbidities, use of systemic therapies, and other potentially confounding factors. Only randomized studies, such as the Eastern Cooperative Oncology Group E2108 trial currently underway, will be able to definitively assess the value of local therapy directed at the primary tumor in this setting.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P4-16-06.

Benefit of radiotherapy dose escalation in localized prostate cancer with respect to expression of intrinsic markers of hypoxia

e16068

Background: Dose escalation improves the efficacy of prostate cancer radiotherapy (RT) at the cost of increased toxicity. Tumor hypoxia causes radioresistance, so the benefit of RT dose escalation may be greater in more hypoxic cancers. Methods: Cases had localized prostate cancer treated with neo-adjuvant androgen deprivation and radical RT at the Royal Marsden in two randomized trials of dose escalation (64 vs 74Gy). Tumour expression of three markers (vascular endothelial growth factor (VEGF), hypoxia inducible factor-1α(HIF-1α), and osteopontin) was assessed immunohistochemically using a semi-quantitative scale by a uro-pathologist, and analyzed with respect to freedom from biochemical failure (FFBF) using the Phoenix definition. Expression of each marker was dichotomised about the median for analysis of the impact of dose-escalation on outcome. Results: 201 cases with a median follow-up of 7 years were evaluable. Seven-year FFBF was 67% vs 40% (HR: 0.42, 95% CI 0.26–0.7, p=0.001) for 74 Gy versus 64Gy, respectively, among cases with high osteopontin expression, and 70% vs 82% (HR: 1.41, 95% CI 0.53–3.76, p=0.49) for 74Gy vs 64Gy among cases with low osteopontin expression. The benefit of RT dose escalation was similar regardless of VEGF or HIF- 1α expression. Conclusions: These data generate the hypothesis that osteopontin expression could inform RT dose individualisation. If validated, patients with low tumor expression of osteopontin could elect to receive less toxic, standard dose RT.

No significant financial relationships to disclose.

CT assessment of conjugate eye deviation in acute stroke

Conjugate eye deviation seen on clinical examination helps to localize pathology in acute ischemic stroke. Eye deviation can also be assessed on a CT head scan. The authors found that CT eye deviation reliably lateralizes to the ischemic hemisphere (positive predictive value 93%) without reference to clinical examination. In an era of thrombolysis and rapid decision making in acute ischemic stroke, eye deviation on CT can help quickly direct attention to the affected hemisphere.

Delayed versus immediate bedside inoculation of culture media for diagnosis of vaginal trichomonosis

A comparison of delayed versus immediate inoculation of culture medium for the diagnosis of trichomonosis was conducted. The sensitivities of the two methods were 100 and 97.4%, respectively. Delayed inoculation of culture medium for women without evidence of trichomonosis on direct microscopic examination is a valid diagnostic procedure.

Validity of self-obtained vaginal specimens for diagnosis of trichomoniasis

A comparison of self- and clinician-collected vaginal specimens for the diagnosis of trichomoniasis was conducted. The sensitivities of culture methods using self- and clinician-collected specimens were 84.6 and 88.5%, respectively. There was no significant difference between the sensitivities of culture methods using self- and clinician-collected vaginal specimens for the diagnosis of trichomoniasis.

The use of sequential self-obtained vaginal smears for detecting changes in the vaginal flora

BACKGROUND AND OBJECTIVES

The ability to study daily changes in the vaginal flora may provide insight into the pathogenesis of bacterial vaginosis. Because culture of the vaginal fluid is tedious and expensive, the utility of self-obtained vaginal smears for documenting changes in the flora was evaluated.

GOALS

To validate the adequacy of self-collected vaginal fluid Gram stains and use them to monitor vaginal flora.

STUDY DESIGN

Ten asymptomatic premenopausal women collected daily vaginal smears for 30 days. The smears were Gram stained and interpreted using a standardized scoring system (Nugent criteria). In addition, results from self- and clinician-obtained vaginal smears from 18 women were compared to validate the adequacy of self-obtained smears.

RESULTS

Two women had asymptomatic bacterial vaginosis. One woman, who was postpartum, had intermediate flora that toward the end of the collection period changed to Lactobacillus predominant. The remaining seven women exhibited two patterns. One was Lactobacillus morphotypes only; the second consisted of Lactobacillus-predominant days interspersed with days with moderate to high numbers of Gardnerella/Bacteroides morphotypes. There was a significant correlation of the point of change in the flora of this group with menses.

CONCLUSIONS

The adequacy of self-collected vaginal fluid Gram's stains was validated. Changes in vaginal flora were demonstrated over a 30-day period by use of this methodology.

Humoral antibody to Mobiluncus curtisii, a potential serological marker for bacterial vaginosis

While bacterial vaginosis (BV) is a polymicrobial syndrome, Mobiluncus spp. are the organisms most highly associated with this condition. It is possible that serum antibody to Mobiluncus spp. could be used as a serological marker for BV. Using immunofluorescence techniques, we studied the prevalence of antibody to M. curtisii among three cohorts-pregnant women, pediatric patients, and sexually inexperienced women. The prevalence of antibody in each of these three groups was 75, 6, and 0%, respectively. Of the three pediatric patients with antibody to Mobiluncus curtisii, two were neonates, and the only class of antibody detected was immunoglobulin G. Among the cohort of pregnant women, the presence of antibody could not be correlated with a clinical history of BV. Serum antibody to M. curtisii could be a useful serological marker for BV. The lack of correlation of antibody positivity to historical information regarding BV suggests that unrecognized or undiagnosed episodes of BV may be common.

Orbital cellulitis and blindness following a blepharoplasty

An unusual case of orbital cellulitis following blepharoplasty, with resultant blindness in that eye, is presented. The cause is unknown, but the pathogenesis and treatment of this rare complication are discussed. Unilateral severe headache may alert one to the possibility of this rare, but grave, complication.