High-grade vaginal intraepithelial neoplasia: can we be selective about who we treat?

Transcription

1 DOI: / Gynaecological oncology High-grade vaginal intraepithelial neoplasia: can we be selective about who we treat? N Ratnavelu, a A Patel, a AD Fisher, a K Galaal, a P Cross, b R Naik a a Northern Gynaecological Oncology Centre, b Pathology Department, Queen Elizabeth Hospital, Gateshead, UK Correspondence: Dr N Ratnavelu, Northern Gynaecological Oncology Centre, Queen Elizabeth Hospital, Gateshead, NE9 6SX, UK. nithya.ratnavelu@ghnt.nhs.uk; nithya_dgr@hotmail.com Accepted 19 February Published Online 2 April Objective To determine the role of conservative management in high-grade vaginal intraepithelial neoplasia (HG VaIN). Design Retrospective observational study. Setting Northern Gynaecological Oncology Centre, Gateshead, UK. Population A total of 100 women with histologically-proven HG VaIN. Methods Review of patient records from 1995 to Main outcome measures Rates of progression to cancer, remission, and disease recurrence, particularly post when vaginoscopy is normal but cytology is abnormal. Results Of 100 women referred, 69 underwent initial of whom 47 (68%) went into remission: of these, seven developed a recurrence after a median follow-up of 29 months (range months). Of the 31 women managed conservatively with cytological and vaginoscopic surveillance, no cancers developed after a median follow-up of 35 months (range months). Rate of overall progression to cancer was 3% and all were detected among the initial group after a median of 59 months (range months). Post-, when normal vaginoscopy was accompanied by abnormal cytology, two categories existed. Of 24 cases with low-grade cytological abnormality, recurrence of HG VaIN occurred in seven (29%) after a median follow-up of 12 months (range months). Of 19 cases with HG cytological abnormality, 15 (79%) developed recurrence at a median follow-up of 7 months (range 2 21 months), giving a hazard ratio 5.6 (95% confidence interval , P = 0.001). Conclusions It is possible to select women with HG VaIN for conservative surveillance with excellent results. The majority of women undergoing initial will enter remission. Post, if cytological abnormality develops in the presence of normal vaginoscopy, the majority of women will develop histological HG VaIN recurrence. Keywords Cytology, high-grade vaginal intraepithelial neoplasia, vaginal intraepithelial neoplasia, vaginoscopy. Please cite this paper as: Ratnavelu N, Patel A, Fisher A, Galaal K, Cross P, Naik R. High grade vaginal intraepithelial neoplasia: can we be selective about who we treat?. BJOG 2013;120: Introduction Vaginal intraepithelial neoplasia (VaIN) is a rare condition, comprising 0.4% of all intraepithelial neoplasia of the lower female genital tract. VaIN has a well-known association with human papillomavirus (HPV) and cervical intraepithelial neoplasia (CIN), seen in up to 43% of women. 1,2 When associated with CIN, the vaginal lesions are likely to be in continuation with the cervical abnormality, 1 in part because the congenital transformation zone extends on to the vagina in 5% of women. 3 The typically upper vaginal location of VaIN 1,4 6 makes accessibility and detection difficult. Given its asymptomatic nature 4,5,7 it is typically detected cytologically, 5,8 in particular after hysterectomy. Due to the complexities in diagnosis and management, and the risk of progression to cancer, 4,7,9 women with highgrade VaIN (grade 2 or 3) are usually referred to specialised tertiary cancer centres for further management. Evidence for the best management of high-grade VaIN is lacking. Current practices include immediate surgical with excision or ablation of all suspicious and/or symptomatic areas 5,10 or medical management (chemoablation). However, repeated surgical s can result in significant surgical and psychosexual morbidity. 5 Conservative surveillance (with vaginoscopy and cytology) for selected populations is also described in the literature 2,4,7 but the efficacy and safety of these approaches is not well known. Deficiencies exist in the majority of the studies: (1) ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG 887

2 Ratnavelu et al. clear definitions for remission and recurrence are lacking; (2) the majority of case series are smaller; (3) those of comparable size date back 10 years 4,7,9 ; (4) previous papers report a mixed series of low-grade and high-grade VaIN; and in particular, (5) the role of abnormal cytology as a preclinical indicator of disease recurrence has not previously been established among women post-. The objective of this study is to establish if we, as clinicians, can be selective of which women we manage surgically and which can be managed with surveillance only. We present here the largest and only specific series of highgrade VaIN described in the literature. Methods Women from northern England with a diagnosis of VaIN have historically been referred to the Northern Gynaecological Oncology Centre (NGOC) based at Gateshead. All such women from January 1995 until June 2011, attending for review at a specific NGOC vaginal precancer clinic, were included. Women were prospectively registered on the NGOC departmental database, from which data were accessed in conjunction with the medical notes. This study of clinical service provision was registered with the Safecare & Audit Department. Women known to have active vaginal cancer at referral and those without histological confirmation of VaIN were excluded. Women s charts and records were reviewed and data were collected regarding demographics, risk factors, management strategy (primary or conservative management), types of (ablative or excisional), remission rate, recurrence rate and progression to cancer. In those women with abnormal cytology but normal vaginoscopy following, recurrence rate of high-grade VaIN was noted. When more than one grade of VaIN was present on histology, women were assigned the highest grade of VaIN. At initial consultation, all women had a vaginoscopic examination by an accredited British Society of Cervical Cytology & Pathology colposcopist, using 3% aqueous acetic acid followed by application of Lugol s iodine. Primary was predefined as before or immediate after referral to the NGOC. Some of these women underwent repeated s. Conservative management was predefined as surveillance with cytology and vaginoscopy in the absence of initial. In general, women were chosen for conservative management, irrespective of referral cytology, if they met the following criteria: no history of genital tract cancer; no evidence of immunosuppression; presence of a unifocal lesion measuring less than 2 cm in diameter; and no vaginoscopic abnormalities suspicious of invasion. Recurrence was defined as histological evidence of VaIN after ; and remission was defined as the absence of recurrence within the first 12 months. Progression was defined as the histological evidence of invasion. Low-grade cytology was predefined as borderline nuclear abnormalities or mild dyskaryosis and high-grade cytology as moderate or severe dyskaryosis or borderline nuclear abnormalities with suspicion of high-grade change. Women who developed abnormal cytology but demonstrated no vaginoscopic abnormality were observed until a vaginoscopic lesion became apparent for biopsy. In the analysis of these, the authors considered each episode when normal vaginoscopy was accompanied by abnormal cytology to be a case. Therefore, each woman could have more than one episode of normal vaginoscopy accompanied by abnormal cytology during her follow-up. The follow-up policy was 3-monthly or 6-monthly surveillance with vaginoscopy and cytology. This became annual review in the presence of normal cytology. When vaginoscopic and cytological findings were both normal over a period of 2 years, women were discharged to primary care if vault or vaginal annual cytological follow-up could be agreed with their general practitioners. Results One hundred women with high-grade VaIN were identified. The median age of women at diagnosis was 50 years (range years). Table 1 shows the indications for referral to the NGOC and clinical risk factors for all women seen. Multifocality of disease was demonstrated in Table 1. Clinical risk factors for all women with high-grade VAIN Clinical risk factors/presentation All women, n = 100 Age Median 50 years (range 22 83) Main referral indication Abnormal cytology/vaginoscopy 66 Histological high grade VaIN on referral 24 Symptoms 8 Non-vaginal precancer/cancer 2 Non-vaginal lower genital tract precancer 56 Antecedent 26 Concurrent 30 Non-vaginal lower genital tract cancer 10 Antecedent 3 Concurrent 7 Smoking history 44 Immunosuppression 4 Previous hysterectomy 48 For CIN 37* For cervical cancer 7 Benign gynaecological 4 *Six had residual CIN in hysterectomy specimen. 888 ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG

3 Management of high grade VaIN. 73% of women and cytological evidence of HPV was reported in 44%. Forty-eight women had a previous hysterectomy, seven of which were for cervical cancer, 37 for CIN and the remaining four for benign gynaecological conditions. These latter four were referred with abnormal cytology, two of which were following diagnosis of VaIN at the vaginal cuff edges on the hysterectomy specimen. The median interval between hysterectomy to development of high-grade VaIN among all women was 2 years (range 0 46 years). Of these 48 post-hysterectomy women, 19 were diagnosed with high-grade VaIN within 1 year of hysterectomy (40%). Seventeen of these 19 had hysterectomy for CIN or cervical cancer and it is unclear whether the development of VaIN in these women was a result of de novo disease or whether this represented continuation or incomplete excision of cervical lesions. High-grade VaIN was located in the upper third of the vagina in 96% of all women who underwent previous hysterectomy. The lesion was described as located in the dog ears, that is the lateral corner recesses of the vaginal vault, in 27% of women. In contrast, among those women who had not had a hysterectomy, CIN was present in only 24 (46%). Of the 100 women referred, 69 underwent primary and 31 were managed conservatively with vaginoscopic and cytological surveillance. The histological outcomes are shown in Figure 1. The rate of progression to cancer was three out of 100 (3%) with all three cancers detected among the primary group at a median of 59 months Remained in remission 40 (85%) Remission 47 (68%) Subsequent recurrence* 7 (15%) 0 Primary 69 Recurrence HG VaIN within 12 months 22 (32%) Subsequent remission 12 (55%) HG VaIN 100 Persisting HG VaIN** 10 (45%) 3 Conservative management 31 No cancer 31 (100%) Subsequent 5 (16%) 0 Figure 1. Histological outcomes of women with high-grade VaIN. *Median recurrence 29 months ( months), repeated performed in seven patients, no cancer. **Repeated performed in seven patients, including three who later developed cancer. (range months) after diagnosis of high-grade disease. All three women had persistent HG VaIN despite multiple surgical excisions. All three women are alive and well at months following diagnosis and of invasive disease and their histories are given as follows. Patient 1, aged 64 years at referral, had previously had a hysterectomy for stage 1a1 squamous cell cancer of cervix but developed high-grade VaIN within 12 months of the hysterectomy. She had primary vaginal excisional for the VaIN but due to residual disease underwent five further surgical procedures and a trial of Imiquimod. Despite this she developed moderately differentiated vaginal squamous cell cancer (stage IV with rectal involvement) at 59 months from VaIN diagnosis, and was treated with radiotherapy and concurrent chemotherapy. Patient 2, aged 37 years at referral, had previously undergone hysterectomy for CIN but developed high-grade VaIN within 12 months of the hysterectomy. She had primary excisional but underwent 15 further surgical procedures and a trial of 5-fluorouracil before developing stage II vaginal cancer at 249 months from diagnosis of high-grade VaIN and was treated with radiotherapy. Patient 3, aged 68 years at referral, had undergone hysterectomy for CIN 19 years previously. She had primary ablative then excisional but developed early stromal invasion at 8 months from diagnosis of high-grade VaIN. Despite three further excisional s and a trial of 5-fluorouracil she developed stage IV vaginal cancer (with bladder involvement) 8 years after diagnosis of early stromal invasion and was treated with radiotherapy. In none of these three women was total vaginectomy performed. Instead directed partial abdominal and vaginal vaginectomies were performed. Of the 69 women managed with primary, 22 had evidence of residual disease or recurrence within 12 months (32%). The remaining 47 women (68%) were in remission. Of those 47 in remission, seven developed a subsequent recurrence after a median follow-up of 29 months (range months). Although the median number of surgical s received by each woman was one (range 1 16), a total of 18 women required repeated with a median of two s (range 2 16). Table 2 shows the modes of for primary and repeated s, and the remission rate among those undergoing primary excision was 33/46 (72%). It is worth noting that 22 women underwent primary ablative. All these were asymptomatic on referral and 18 of the 22 had vaginsocopic evidence of multifocal disease. Excisional was favoured for the majority of women undergoing repeated s. This was because of concern regarding increased risk of invasion with recurrent VaIN. Of the 31 women managed conservatively, five required surgical at a median of 16 months (range 9 ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG 889

4 Ratnavelu et al. Table 2. Methods of for high grade VaIN Method of Primary, n = 69 Repeated n = 18* Treatment following conservative management, n = 5 Medical Ablative Excisional Abdominal Vaginal *18 women underwent one or more repeated s. 24 months) from first review, after which one woman had a recurrence. However, none progressed to cancer after a median period of 35 months (range months) from diagnosis resulting in a total follow-up of 121 womanyears. As demonstrated in Figure 2, the authors analysed histological outcomes of women who had previously been treated but developed subsequent abnormal cytology in the presence of normal vaginoscopy. Forty-three such cases occurred in 31 women post-. These were analysed in two groups. The first group comprised 24 cases (56%) where low-grade cytology was accompanied with normal vaginoscopy. Seven (29%) of these 24 cases developed subsequent high-grade VaIN after a median of 12 months (range months) and none progressed to cancer. The remaining 17 cases remained recurrence-free after a median follow-up of 27 months (range 5 76 months). The second group comprised 19 cases (44%) where high-grade cytology was accompanied with normal vaginoscopy. Fifteen (79%) developed recurrence of high-grade VaIN at a median of 7 months (range 2 21 months). Kaplan Meier analysis of cumulative hazard function demonstrates a statistically significant difference (P < 0.001) in high-grade VaIN recurrence rate between those with lowgrade and high-grade cytology post- (Figure 3). Hazard for recurrence of high-grade VaIN with high-grade cytology, compared with low-grade cytology, was 5.6 (95% confidence interval , P = 0.001). The median number of clinic appointments was seven (range 1 70 visits). Fifty-eight women were discharged, including 13 who migrated or self-discharged, at a median length of follow-up of 35 months (range months) from diagnosis. Discussion Main findings Our results support the view that, with appropriate selection of women, conservative surveillance with vaginoscopy and cytology is safe and equally efficient in the management of a significant proportion of women with high-grade VaIN. Among the cohort of women managed conservatively no cancers developed. This study also adds valuable information regarding the role of cytological surveillance after. We have Abnormal cytology Normal vaginoscopy 43 cases LG cytology 24 (56%) HG cytology 19 (44%) Remission 17 (71%) HG VaIN recurrence 7 (29%) Remission 4 (21%) HG VaIN recurrence 15 (79%) 1 Figure 2. Outcomes of abnormal cytology with normal colposcopy. LG, low grade; HG, high grade; VaIN, vaginal intraepithelial neoplasia. Figure 3. Kaplan Meier analysis of cumulative hazard function to develop high-grade VaIN among cases with abnormal cytology but normal vaginoscopy post-. Log rank test P < ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG

5 Management of high grade VaIN. demonstrated that in the presence of low-grade cytology, there is a 29% rate of high-grade VaIN recurrence and so there is potential for a conservative approach to these women s disease, especially as none progressed to cancer. However, those with high-grade cytology should be monitored more frequently as there is a 79% rate of high-grade recurrence within the first 2 years after. Strengths This study describes the largest cohort of women with histologically proven high-grade VaIN. We used predefined outcome measures to standardise assessment of response to. This study benefits from presenting the outcomes of a single institution, wherein the management policies have remained relatively unchanged. Our study benefits from long-term follow-up data and therefore the authors have included a flowchart demonstrating a management plan for this uncommon premalignant condition as given in Figure 4. Furthermore, this study is the first of its kind to examine the utility of cytological surveillance in the absence of vaginoscopic abnormalities in predicting the recurrence of high-grade VaIN. YES Conservative management If both vaginoscopy & cytology normal over 2-year period Discharge to GP with agreement for annual cytology If abnormal cytology: re-refer Clinical/cytological/histological suspicion of HG VaIN 1. Biopsy and documentationof lesion characteristics* 2. Refer to tertiary centre Tertiary centre: 1. Documentation of lesion characteristics 2. Biopsy if not performed 3. Assess symptoms & discuss options with patient Criteria met irrespective of referral cytology : 1. No history of genital tract cancer 2. Not immunosuppressed 3. Unifocal lesion 4. Less than 2-cm diameter 5. Vaginoscopic findings not suspicious of invasion Follow-up vulvovaginal clinic every 3 6 months: cytology, vaginoscopy +/ colposcopy +/ biopsy Primary : consider ablation/excision If biopsy-confirmed high grade VaIN Figure 4. Flowchart showing recommended management of highgrade VaIN. LG, low grade; HG, high grade; VaIN, vaginal intraepithelial neoplasia. *Size, density of acetowhite/iodine negativity, focality. NO Weaknesses Given the retrospective nature of this study, we would recommend prospective study of our selection criteria to support our findings. It is our understanding that clinical risk assessment based on overall history, in particular presence of risk factors (such as history of previous lower genital cancer and immunosuppression) and clinical assessment of lesion (multifocality, large lesion 2 cm, and suspicious findings on vaginoscopy) played a role in such decisionmaking on the management approach. We believe that expertise can be developed with experience and adequate caseload to select women with high-grade VaIN who will benefit from a surveillance-only strategy. Ablative has not been well-represented in this study. Given the importance of histological diagnosis we have preferred to opt for excisional rather than ablative surgical management. Interpretation The age at diagnosis 2,10 12 and patient risk factors 1,3,4,13 presented here are comparable with what is described in the literature. The low overall progression to cancer of 3% is also in keeping with other studies, reporting rates as high as 12%. 2,5,8 It is natural to expect rates of invasion to increase with duration of follow-up and women with risk factors for invasive disease should be flagged, or highlighted, at initial referral. 5,14 It is worth noting that the three women who developed invasive disease in our study demonstrated persistent residual high-grade VaIN following each. In keeping with the field carcinogenesis theory previously described by Marcus and others, post-hysterectomy for CIN almost all women had disease located in the upper third of the vagina, as previously described in the literature. 10,11,13,14 Therefore, preoperative evaluation of the vaginal mucosa before definitive surgery for cervical dysplasia or cancer is vital, 13 particularly the vaginal vault recesses, which can be common sites of recurrence. Our study underlines the importance of post-hysterectomy vaginoscopic and cytological follow-up of these vault recesses and is in keeping with published literature reporting that between 0.9 and 6.8% of women undergoing hysterectomy for CIN develop VaIN. 4,11,13,18 20 Conversely, the risk of VaIN after hysterectomy for benign conditions is 1.3% at 10 years 5,21 and no agreed cytological follow-up has been established or is recommended in this setting. There may well be an association between HPV subtype, remission rate and grade of VaIN. Our study describes a 68% remission rate of high-grade VaIN, in particular after a single excisional (72%). This is similar to other studies reporting rates of 35 67%. 5,11 The heterogeneity of HPV subtypes seen in vaginal epithelium may possibly ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG 891

6 Ratnavelu et al. explain these high remission rates, 6,22 in that although around 90% of VaIN is associated with HPV, 22 only half of these are associated with the high-risk HPV ,24 The likelihood that biopsy alone accounts for such a high remission is unlikely because the majority of cases of VaIN are multifocal. Frega et al. 25 showed that positivity to HPV-16/18 was significantly higher in women with recurrent VaIN than those who remained in remission. Future studies In future, it may be possible that HPV-16/18 testing after could help triage women into two groups: those who are HPV-16/18-positive who could be surveyed every 3 months; and those with non-16/18 subtypes requiring less frequent surveillance. Inclusion of HPV testing in addition to cytological surveillance may aid prediction of high-grade recurrence before cytological or vaginoscopic abnormality. Future studies should also include prospective cohort studies to not only aid prediction of the course of the disease, in particular recurrence and progression, but also response to. Such studies may provide objective estimates of probability of recurrence and invasion as a supplement to clinical assessment. 26 Conclusions In summary, this study demonstrates the safety and efficacy of conservative management in appropriately selected cases of high-grade VaIN and highlights the importance of cytological and vaginoscopic follow-up. We believe that employing the management strategy suggested can further improve monitoring of this rare condition and aid in selection of women for conservative management. Disclosure of interests All authors declare there are no conflicts of interest. Contribution to authorship RN and PC devised the study, supervised the data collection and contributed to the discussions. NR and PC collected the data. NR, AP and AF performed the analyses. All authors contributed to the discussions and approved the final manuscript. Details of ethics approval This audit of clinical service provision was prospectively registered with the Safecare & Audit Department and ethics approval was not necessary. Funding None. Acknowledgements The authors would like to thank Mr Daniel Spelman for collecting the computerised data and Dr Ioannis Biliatis for proof-reading. There are no conflicts of interest. & References 1 Nwabineli N, Monaghan J. Vaginal epithelial abnormalities in patients with CIN: clinical and pathological feautures and management. Br J Obstet Gynaecol 1991;98: Aho M, Vesterinen E, Meyer B, Purola E, Paavonen J. Natural history of vaginal intraepithelial neoplasia. 1991;68: Woodman C, Jordan J, Wade-Evans T. The management of vaginal intraepithelial neoplasia after hysterectomy. BJOG 1984;91: Dodge J, Eltabbakh G, Mount S, Walker R, Morgan A. Clinical features and risk of recurrence among patients with vaginal intraepithelial neoplasia. Gynecol Oncol 2001;83: Sillman FH, Fruchter RG, Chen YS, Camillien L, Sedlis A, McTuqye E. Vaginal intraepithelial neoplasia: risk factors for persistence, recurrence, and invasion and its management. Am J Obstet Gynecol 1997;176: Massad S. Outcomes after diagnosis of vaginal intraepithelial neoplasia. J Low Genit Tract Dis 2008;12: Diakomanolis E, Stefanidis K, Rodolakis A, Haidopoulos D, Sindos M, Chatzipappas I, et al. Vaginal intraepithelial neoplasia: report of 102 cases. Eur J Gynaecol Oncol 2002;23: Sherman M, Paull G. Vaginal intraepithelial neoplasia: reproducibility of pathologic diagnosis and correlation of smears and biopsies. Acta Cytol 1993;37: Rome R, England P. Management of vaginal intraepithelial neoplasia: a series of 132 cases with long-term follow-up. Int J Gynecol 2000;10: Lenehan P, Meffe F, Lickrish G. Vaginal intraepithelial neoplasia: biologic aspects and management. Obstet Gynecol 1986;68: Benedet J, Boyes A, Nicols T, Millner A. The role of colposcopy in the evaluation of abnormal vaginal vault smears. Gynecol Oncol 1977;5: Murta E, Neves Junior M, Sempionato L, Costa M, Maluf P. Vaginal intraepithelial neoplasia: clinical-therapeutic analysis of 33 cases. Arch Gynecol Obstet 2005;272: Jimerson G, Merrill J. and dysplasia of the post-hysterectomy vaginal cuff. Gynecol Oncol 1976;4: Terzakis E, Androutsopoulos G, Zygouris D, Grigoriadis C, Derdelis G, Argiannaki N. Loop electrosurgical excision procedure in Greek patients with vaginal intraepithelial neoplasia. Eur J Gynaecol Oncol 2010;31: Marcus S. Multiple squamous cell carcinomas involving the cervix, vagina and vulva: the theory of multicentric origin. Am J Obstet Gynecol 1960;80: Newman W, Cromer J. The multicentric origin of carcinomas of the female anogenital tract. Surg Gynecol Obstet 1959; Hammond I. Multicentric carcinoma of the female lower genital tract. Br J Obstet Gynaecol 1983;90: Ireland D, Monaghan J. The management of the patient with abnormal vaginal cytology following hysterectomy. Br J Obstet Gynaecol 1988;95: Colposcopy and Programme Management: Guidelines for the NHS Cervical Screening Programme, 2nd edn. London: NHSCSP Publication No 20. May Gemmell J, Holmes D, Dunca I. How frequently need vaginal smears to be taken after hysterectomy for cervical intraepithelial neoplasia? Br J Obstet Gynaecol 1990;97: ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG

7 Management of high grade VaIN. 21 Pearce KF, Haefner HK, Sarwar SF, Nolan TE. Cytopathological findings on vaginal Papinicolaou smears after hysterectomy for benign gynecologic disease. N Engl J Med 1996;335: Sugase M, Matsukura T. Distinct manifestations of human papillomaviruses in the vagina. Int J 1997;72: Daling J, Madeleine M, Schwartz S, Shera K, Carter J, McKnight B, et al. A population-based study of squamous cell vaginal cancer: HPV and cofactors. Gynecol Oncol 2002;84: Srodon M, Stoler M, Baber G, Kurman R. The distribution of low and high-risk HPV types in vulvar and vaginal intraepithelial neoplasia (VIN and VAIN). Am J Surg Pathol 2006;30: Frega A, French D, Cerekja A, Vetrano G, Moscarini M. Prediction of persistent vaginal intraepithelial neoplasia in previously hysterectomized women by high-risk HPV DNA detection. Lett 2007;249: Moons K, Royston P, Vergouwe Y, Grobbee D, Altman D. Prognosis and prognostic research: what, why, and how? BMJ 2009;338: ª 2013 The Authors BJOG An International Journal of Obstetrics and Gynaecology ª 2013 RCOG 893