Advertisement

European Radiology

, Volume 29, Issue 10, pp 5717–5722 | Cite as

Comparison of 3D endoanal ultrasound and external phased array magnetic resonance imaging in the diagnosis of obstetric anal sphincter injuries

  • Jaan KirssJrEmail author
  • Heikki Huhtinen
  • Eini Niskanen
  • Jyrki Ruohonen
  • Marja Kallio-Packalen
  • Sarita Victorzon
  • Mikael Victorzon
  • Tarja Pinta
Open Access
Gastrointestinal
  • 474 Downloads

Abstract

Objectives

The gold standard of postpartum anal sphincter imaging has been the 3D endoanal ultrasound (EAUS). Development of magnetic resonance imaging (MRI) has allowed anal sphincter evaluation without the use of endoanal coils. The aim of this study is to compare these two modalities in diagnosing residual sphincter lesions post obstetric anal sphincter injury (OASI).

Methods

Forty women were followed up after primary repair of OASI with both 3D EAUS and external phased array MRI. Details of the anal sphincter injury and sphincter musculature were gathered and analysed.

Results

There was a moderate interrater reliability (κ = 0.510) between the two imaging modalities in detecting sphincter lesions, with more lesions detected by MRI. There was a moderate intraclass correlation (ICC) between the circumference of the tear (κ = 0.506) and a fair ICC between the external anal sphincter thickness measurements at locations 3 and 9 on the proctologic clock face (κ = 0.320) and (κ = 0.336).

Conclusions

The results of our study indicate that the use of external phased array MRI is feasible for detecting obstetric anal sphincter lesions postpartum. This allows for imaging of the sphincter defects in centres where EAUS imaging is not available.

Key Points

• A two centre prospective study that showed external phased array MRI to be a valid imaging modality for diagnosing obstetric anal sphincter injuries.

Keywords

Magnetic resonance imaging Ultrasonography Anal canal Rupture Postpartum period 

Abbreviations

3D EAUS

Three-dimensional endoanal ultrasound

EAUS

Endoanal ultrasound

FI

Faecal incontinence

MRI

Magnetic resonance imaging

OASI

Obstetric anal sphincter injury

Introduction

The main aetiological factor associated in developing faecal incontinence (FI) is vaginal birth. The aetiology of postpartum FI is multifactorial, with injury to the anal sphincters as the principal cause. Injury to the pudendal nerve, puborectalis muscle, the anal sphincter complex, or the combination of these factors could also play a role in developing postpartum FI [1, 2, 3].

Even after a successful initial repair of the sphincter lesion, women with obstetric anal sphincter injury (OASI) have a 50% greater risk of developing faecal incontinence compared to patients without such an injury [4, 5]. In addition to FI, a history of OASI can have a negative effect on sexual function. It has been reported that women with a history of OASI are less likely to plan future pregnancies [6, 7, 8].

It is estimated that 1–5% of all vaginal deliveries complicate in perineal tears worldwide [9, 10, 11]. There were 53,614 births registered in Finland in 2016; 1.2% of these were complicated by a grade III or IV perineal tear [10].

Vaginal tears are generally repaired by a midwife immediately postpartum. When a sphincter lesion is suspected, the repair is conducted by a gynaecologist or a gastrointestinal (GI) surgeon in operating room conditions with proper anaesthesia [12].

Perineal tears are graded from 1 to 4, according to Sultan et al [13] (Table 1). Grade 3 and 4 perineal tears can also be classified as obstetric anal sphincter injuries (Fig. 1). When a grade 3–4 perineal tear is missed or the initial repair was not successful, the results of secondary repair are usually not encouraging [14, 15, 16]. It has been shown that withholding repair until operating room facilities and experienced personnel are available improves the outcomes of primary repair. A delay of up to 24 h has not shown to have a negative effect on the outcomes of primary sphincteroplasty (SP) [17, 18].
Table 1

Grading of obstetric injuries according to Sultan et al

Grade 1

Superficial tear of the vaginal mucosa

Grade 2

Tearing of the vaginal mucosa and perineal muscles

Grade 3a

Tear of the EAS, < 50% of the muscle thickness involved

Grade 3b

Tear of the EAS, > 50% of the muscle thickness involved

Grade 3c

Complete EAS and IAS tear

Grade 4

Tear involving the rectal mucosa

Fig. 1

Schematic representation of grade 3–4 perineal tears: (a) grade 3a tear, (b) grade 3b tear, (c) grade 3c tear, (d) grade 4 tear

Even without an OASI, some women experience a certain degree of FI postpartum, which often subsides after a couple of weeks. This makes the diagnosis of residual OASI challenging without proper imaging.

Endoanal ultrasound (EAUS) has been the gold standard for diagnosing sphincter lesions [19]. Though it is an inexpensive and safe imaging modality, it is operator-specific and not widely available in non-specialist centres. Studies where the results of EAUS and endoanal magnetic resonance imaging (MRI) have been compared have shown that the two modalities are in fact comparable in evaluating sphincter defects [20]. The widespread use of MRI and ever-improving quality of these images has opened up the possibility for imaging the sphincter musculature without the use of endoluminal coils. This has particular clinical value in Finland, where EAUS imaging is available in only 2 of the 16 central hospitals and 3 of the 5 university hospitals, whereas MRI is available in all central and university hospitals.

The aim of this study is to compare 3D EAUS imaging with MRI in the diagnosis of OASI and determine whether external phased array pelvic MRI at 1.5 T is suitable for the diagnosis of OASI.

Materials and methods

This is a prospective observational study conducted in two Finnish Central hospitals. All women delivering in Vaasa and Seinäjoki Central hospitals between 01/2014 and 08/2017 who had sustained a third or fourth degree OASI diagnosed immediately postpartum were informed of the study. Women suspected of having a missed OASI postpartum were also invited to participate in the study. Written consent was obtained from all participants. All participants were asked to fill out the Wexner incontinence questionnaire. Demographic, obstetric, and follow-up data was collected from electronic patient archives.

All women consenting to participate underwent 1.5 T MR imaging of the rectal musculature 3–12 months after childbirth. Participants were called for a follow-up visit to the surgical outpatient clinic after undergoing MR imaging, 8–12 months postpartum. Upon the follow-up visit, a 3D endoanal ultrasound study was conducted, and patients were informed of the MRI and 3D EAUS study results. All women with symptomatic FI were referred to a physiotherapist for biofeedback and pelvic floor physiotherapy. Women who were still symptomatic 6–8 weeks after starting physical therapy were evaluated for possible future surgical treatment.

The B-K Medical® Ultrasound Scanner Type 2202 with a Type 2050H endoanal probe was used for ultrasound imaging. The imaging was performed by a GI surgeon with the patient in the left lateral position. Later the images were reviewed by two GI surgeons together and analysed for sphincter thickness and scale of sphincter tear, if present.

The MRI devices used were the 1.5 T Discovery MR450, GE Healthcare at Vaasa Central hospital and the Siemens® MAGNETOM® Avanto 1.5 T (software version syngo MR B19) at Seinäjoki Hospital. The MRI procedure was performed with the patient in the resting supine position with the 12-channel body matrix external phased array coil. No intravenous contrast agent, nor rectal or anal contrasting, was used. The details of the MRI procedure are presented in Appendices 1 and 2. The MR images were analysed by two radiologists for the same parameters as the 3D EAUS images. Both of the radiologists were blinded to the 3D EAUS result.

A pilot study was performed that indicated a correlation coefficient (r) of 0.788 between the two imaging studies in detecting sphincter lesions. The power analysis revealed that data of 40 women would be needed to prove a correlation (r = 0.6) between MRI and 3D EAUS findings with a power of 0.9 (90%) [21].

Patient data was analysed using Microsoft® Excel for MAC version 15.13.1, IBM® SPSS® software Version 23, and SAS System Version 9.4. The intraclass correlation (ICC) values were calculated for continuous variables and interrater reliability (IRR) for categorical variables. The κ values from 0.0 to 0.2 indicated a slight agreement, values from 0.21 to 0.40 a fair agreement, values from 0.41 to 0.60 a moderate agreement, values from 0.61 to 0.80 a substantial agreement, and values from 0.81 to 1.0 a perfect agreement [22]. Speakman’s correlation coefficient was used to calculate correlations between ordinal data.

This study was approved by the Hospital District of Southwest Finland Ethics Committee (ETMK 66/1801/2015). The study was registered with Clinicaltrials.gov, identification number NCT 03039374.

Results

There were 40 women who consented to participate in the study. The study period was from October 2014 to September 2017. The mean age of women was 29.97 years (SD 4.386), with a mean BMI of 24.82 kg/m2 (SD 4.729). Most of the women were primiparous (n = 25; 62.5%). The median time of birth was at 40 + 2 (range from 37 + 0 to 42 + 6) weeks. Spontaneous births occurred in 52.5% (n = 21) cases; in all other cases, the births were vacuum assisted (n = 19). Both breech position and shoulder dystocia occurred in one case. Most babies presented in either sinciput (n = 16; 40%) or vertex (n = 14; 35%) position. The mean time from partum to sphincter repair was 3 h 40 min (SD: 4 h 49 min). The majority of OASIs were repaired by a gynaecologist (n = 25; 62.5%); 12 (30.0%) of the tears were repaired by a gastrointestinal surgeon and 3 (7.5%) by a midwife.

Women with an OASI spent an average of 3.35 days (SD: 0.862 days) in hospital postpartum.

The average time from delivery to the MRI study was 235.8 days (SD 138.16 days) and 211.27 days (SD 145.9 days) to the 3D EAUS study (p < 0.001).

Of the 40 women analysed, 4 underwent secondary SP performed by a gastrointestinal surgeon. None of the patients were considered for sacral nerve stimulation implantation.

There were 13 EAS tears detected by 3D EAUS and 15 by MRI. The details of the imaging results are presented in Tables 2 and 3. There was a moderate interrater reliability (κ = 0.510; p < 0.001) between the two imaging modalities when determining types of tears. The inter-class coefficient showed a moderate agreement in determining the circumference of the tear (κ = 0.506). There was only a fair ICC between the thickness of the EAS measured at position 3 (κ = 0.320) and 9 (κ = 0.336) on the proctologic clock face.
Table 2

Types of sphincter injuries detected with EAUS and MRI

  

n

%

Kappa1

p value2

3D EAUS finding

No tear

13

32.5

0.510

< 0.001

IAS tear

1

2.5

EAS tear

13

32.5

IAS&EAS tear

13

32.5

MRI finding

No tear

10

25.0

IAS tear

2

5.0

EAS tear

15

37.5

IAS&EAS tear

13

32.5

1Interrater reliability

2Significance of the agreement (McNemar test)

Table 3

MRI and 3D EAUS findings of the circumference of the sphincter tear and EAS thickness measurements at positions 9 and 3 on the proctologic clock face

 

Mean

Std. deviation

Kappa (ICC)1

Mean MRI tear in degrees

63.75

51.375

0.506

Mean 3D EAUS tear in degrees

48.11

38.901

Mean MRI EAS thickness at 92 (mm)

2.693

0.786

0.336

Mean 3D EAUS EAS thickness at 9 (mm)

2.698

0.794

Mean MRI EAS thickness at 33 (mm)

2.753

1.078

0.320

Mean 3D EAUS EAS thickness at 3 (mm)

2.838

0.762

1Intraclass coefficient values

2Position 9 on the proctologic clock face

3Position 3 on the proctologic clock face

The median Wexner score on follow-up was 3 (range 0–11), with a mean score of 4.07 (SD = 3.198).

Discussion

3D EAUS has been the gold standard for the diagnosis of OASI [23]. 3D EAUS hardware is not widely available in Finland, unlike MR imaging, which is available in all centres that provide obstetric services. There are only five centres in Finland were imaging of the sphincters is possible with 3D EAUS. Women delivering in centres with no 3D EAUS imaging capability were not followed up with imaging modality regardless of having FI symptoms postpartum. Results of this study will enable physicians to conduct imaging of the anal sphincter complex in the absence of EAUS hardware. Studies have shown that mere palpation of the anal sphincters is not accurate enough to diagnose a sphincter lesion [24, 25].

Although there have been studies evaluating the correlation of EAUS with endoanal MR imaging on diagnosing sphincter lesions [20], there have been no studies comparing external phased array MRI and 3D EAUS imaging in diagnosing postpartum OASI. Studies where EAUS imaging is compared to endoanal MRI show that although EAUS is superior in diagnosing IAS defects, endoanal MRI is as sensitive in diagnosing EAS defects [20, 26, 27].

The results of our study indicate that there is a moderate interrater reliability between the MRI and EAUS results in detecting IAS and EAS lesions (Table 2). This is in accordance with previously published research on the comparison of EAUS and endoanal MRI findings [26, 27, 28]. In our study, the external phased array MRI was in fact more sensitive, detecting more EAS defects than the EAUS. This is possibly due to the fact that EAUS is unable to differentiate between muscle fibres and scar tissue. In the cases where there was no tear detected on 3D EAUS, the MRI revealed an extensive defect that was replaced by scar tissue. Studies have indicated that extensive scarring in the EAS interferes with the contraction of the EAS, owing to symptoms of FI [29]. It can be argued that the use of MRI could lead to the overdiagnosis of OASI [30]. Even if this was the case, it would not affect the number of patients being treated. It is the view of the authors that SP should be considered for symptomatic patients, no later than 2 years after the initial injury.

There was a moderate correlation between the two imaging modalities in the measurements of EAS defect circumference. This was expected, since there were no probes, etc. in the anal canal during the MR imaging. This made the evaluation of the defect circumference challenging in the MR images. The fair correlation between EAS sphincter thickness measurements was due to the small size of the structures evaluated. The thickness of the female EAS is 2–3 mm [31, 32]. This makes the measurement of these structures with a 3-mm slice thickness inherently imprecise (Fig. 2). Smaller slice thickness and use of 3D MRI could improve this result. Since the thickness of the EAS has no clinical relevance, this finding should not influence possible treatment strategies.
Fig. 2

Image of an anterior EAS defect. a EAUS image. b MR image. The defect is marked with a white arrow

Young patients with symptomatic OASI should be evaluated for possible sphincter repair before SNS treatment. Results from previously published studies indicate that EAS repair conducted on young patients has a higher level of success [16, 33]. Surgical repair is especially important in patients with a very low perineum or a rectovaginal fistula. Since surgical repair is only available for EAS defects, an external phased array MRI at 1.5 T is sufficient in diagnosing these defects and thus planning surgical treatment for symptomatic patients.

The main limitation of this study was the lack of a control group of healthy women. In addition, the study was performed using only 1.5-T MRI scanners. A 3-T scanner could increase the signal-to-noise-ratio of MR images thus possibly improving the detection of abnormalities. However, since 1.5-T scanners are still more common and more easily available, this study was conducted using only 1.5-T scanners.

Our results indicate that external phased array MRI at 1.5 T is suitable for evaluating the anal sphincter complex postpartum. In a setting where EAUS imaging is not accessible, external phased array MRI can be used to evaluate possible residual sphincter injuries.

Notes

Acknowledgements

Open access funding provided by University of Turku (UTU) including Turku University Central Hospital. The authors would like to thank Ms. Elizabeth Prouty, for her revision and corrections of the English language.

The authors would also like to express their sadness due to the passing of a friend and an author of this paper—professor Mikael Victorzon.

Funding

This study has received funding by Finnish Government Research Grant (grant number: 101026).

Compliance with ethical standards

Guarantor

The scientific guarantor of this publication is Tarja Pinta, MD, PhD.

Conflict of interest

The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.

Statistics and biometry

Saija Hurme, MSc, from Turku Clinical Research Centre Department of Biostatistics kindly provided statistical advice for this manuscript.

Informed consent

Written informed consent was obtained from all subjects (patients) in this study.

Ethical approval

Institutional Review Board approval was obtained.

Methodology

• prospective

• observational

• multicentre study

Supplementary material

330_2019_6125_MOESM1_ESM.docx (15 kb)
ESM 1 (DOCX 15 kb)

References

  1. 1.
    Dudding T, Vaizey C, Kamm M (2008) Obstetric anal sphincter injury: incidence, risk factors, and management. Ann Surg 247:224–237CrossRefPubMedGoogle Scholar
  2. 2.
    Aitola P, Lehto K, Fonsell R, Huhtala H (2010) Prevalence of faecal incontinence in adults aged 30 years or more in general population. Colorectal Dis 12:687–691Google Scholar
  3. 3.
    Cerro CR, Franco EM, Santoro GA, Palau MJ, Wieczorek P, Espuña-Pons M (2017) Residual defects after repair of obstetric anal sphincter injuries and pelvic floor muscle strength are related to anal incontinence symptoms. Int Urogynecol J 28:455–460Google Scholar
  4. 4.
    McNicol FJ, Bruce CA, Chaudhri S et al (2010) Management of obstetric anal sphincter injuries--a role for the colorectal surgeon. Colorectal Dis 12:927–930Google Scholar
  5. 5.
    Sundquist JC (2012) Long-term outcome after obstetric injury: a retrospective study. Acta Obstet Gynecol Scand 91:715–718Google Scholar
  6. 6.
    Fodstad K, Staff AC, Laine K (2016) Sexual activity and dyspareunia the first year postpartum in relation to degree of perineal trauma. Int Urogynecol J 27:1513–1523CrossRefPubMedGoogle Scholar
  7. 7.
    Cornelisse S, Arendsen LP, van Kuijk SM, Kluivers KB, van Dillen J, Weemhoff M (2016) Obstetric anal sphincter injury: a follow-up questionnaire study on longer-term outcomes. Int Urogynecol J 27:1591–1596CrossRefPubMedGoogle Scholar
  8. 8.
    Kumar R, Ooi C, Nicoll A (2012) Anal incontinence and quality of life following obstetric anal sphincter injury. Arch Gynecol Obstet 285:591–597CrossRefPubMedGoogle Scholar
  9. 9.
    Smith LA, Price N, Simonite V, Burns EE (2013) Incidence of and risk factors for perineal trauma: a prospective observational study. BMC Pregnancy Childbirth 13:59CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Heino A, Vuori E, Gissler M (2017) Perinatal statistics – parturients, deliveries and newborns 2016. Statistical report, pp 1798–0887Google Scholar
  11. 11.
    Marsh F, Lynne R, Christine L, Alison W (2011) Obstetric anal sphincter injury in the UK and its effect on bowel, bladder and sexual function. Eur J Obstet Gynecol Reprod Biol 154:223–227CrossRefPubMedGoogle Scholar
  12. 12.
    Aigmueller T, Bader W, Beilecke K et al (2015) Management of 3rd and 4th degree perineal tears after vaginal birth. German Guideline of the German Society of Gynecology and Obstetrics (AWMF Registry No. 015/079, October 2014). Geburtshilfe Frauenheilkd 75:137–144CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    Sultan AH (1999) Editorial: obstetrical perineal injury and anal incontinence. Clinical Risk 5:193–196Google Scholar
  14. 14.
    Zorcolo L, Covotta L, Bartolo DC (2005) Outcome of anterior sphincter repair for obstetric injury: comparison of early and late results. Dis Colon Rectum 48:524–531Google Scholar
  15. 15.
    Barisic GI, Krivokapic ZV, Markovic VA, Popovic MA (2006) Outcome of overlapping anal sphincter repair after 3 months and after a mean of 80 months. Int J Colorectal Dis 21:52–56Google Scholar
  16. 16.
    Pinta T, Kylänpää-Bäck ML, Salmi T, Järvinen HJ, Luukkonen P (2001) Delayed sphincter repair for obstetric ruptures: analysis of failure. Colorectal Dis 5:73–78Google Scholar
  17. 17.
    Kirss J, Pinta T, Victorzon S, Kallio-Packalén M, Victorzon M (2016) External phased-array magnetic resonance imaging in the diagnosis of obstetric anal sphincter injury- a pilot study. Turku University Hospital, ESCP 11th Scientific meeting, MilanGoogle Scholar
  18. 18.
    Nordenstam J, Mellgren A, Altman D et al (2008) Immediate or delayed repair of obstetric anal sphincter tears-a randomised controlled trial. BJOG 115:857–865CrossRefPubMedGoogle Scholar
  19. 19.
    Sultan AH, Kamm MA, Talbot IC, Nicholls RJ, Bartram CI (1994) Anal endosonography for idenwying external sphincter defects confirmed histologically. Br J Surg 81:463–465Google Scholar
  20. 20.
    Tan E, Anstee A, Koh DM, Gedroyc W, Tekkis PP (2008) Diagnostic precision of endoanal MRI in the detection of anal sphincter pathology: a meta-analysis. Int J Colorectal Dis 23:641–651Google Scholar
  21. 21.
    Kirss J, Pinta T, Böckelman C, Victorzon M (2016) Factors predicting a failed primary repair of obstetric anal sphincter injury. Acta Obstet Gynecol Scand 95:1063–1069Google Scholar
  22. 22.
    Hallgren KA (2012) Computing inter-rater reliability for observational data: an overview and tutorial. Tutor Quant Methods Psychol 8:23–34CrossRefPubMedPubMedCentralGoogle Scholar
  23. 23.
    Albuquerque A (2015) Endoanal ultrasonography in fecal incontinence: current and future perspectives. World J Gastrointest Endosc 7:575CrossRefPubMedPubMedCentralGoogle Scholar
  24. 24.
    Jeppson PC, Paraiso MF, Jelovsek JE, Barber MD (2012) Accuracy of the digital anal examination in women with fecal incontinence. Int Urogynecol J 23:765–768CrossRefPubMedGoogle Scholar
  25. 25.
    Dobben AC, Terra MP, Deutekom M et al (2006) Anal inspection and digital rectal examination compared to anorectal physiology tests and endoanal ultrasonography in evaluating fecal incontinence. Int J Colorectal Dis 22:783–790Google Scholar
  26. 26.
    West R, Dwarkasing S, Briel JW et al (2005) Can three-dimensional endoanal ultrasonography detect external anal sphincter atrophy? A comparison with endoanal magnetic resonance imaging. Int J Colorectal Dis 20:328–333Google Scholar
  27. 27.
    Dobben AC, Terra JP, Frederik JF et al (2007) External anal sphincter defects in patients with fecal incontinence: comparison of endoanal MR imaging and endoanal US. Radiology 242:463–471Google Scholar
  28. 28.
    Malouf AJ, Williams AB, Halligan S, Bartram CI, Dhillon S, Kamm MA (2000) Prospective assessment of accuracy of endoanal MR imaging and endosonography in patients with fecal incontinence. AJR Am J Roentgenol 175:741–745Google Scholar
  29. 29.
    Terra MP, Beets-Tan RG, van der Hulst VP et al (2006) MRI in evaluating atrophy of the external anal sphincter in patients with fecal incontinence. AJR Am J Roentgenol 187:991–999Google Scholar
  30. 30.
    Thomas GP, Gould LE, Casunuran F, Kumar DA (2017) A retrospective review of 1495 patients with obstetric anal sphincter injuries referred for assessment of function and endoanal ultrasonography. Int J Colorectal Dis 32:1321–1325Google Scholar
  31. 31.
    Soerensen MM, Pedersen BG, Santoro GA, Buntzen S, Bek K, Laurberg S (2014) Long-term function and morphology of the anal sphincters and the pelvic floor after primary repair of obstetric anal sphincter injury. Colorectal Dis 16:O347–O355Google Scholar
  32. 32.
    Tirumanisetty P, Prichard D, Fletcher JG, Chakraborty S, Zinsmeister AR, Bharucha AE (2018) Normal values for assessment of anal sphincter morphology, anorectal motion, and pelvic organ prolapse with MRI in healthy women. Neurogastroenterol Motil 30:e13314CrossRefPubMedPubMedCentralGoogle Scholar
  33. 33.
    Wald A, Bharucha AE, Cosman BC, Whitehead WE (2014) ACG clinical guideline: management of benign anorectal disorders. Am J Gastroenterol 109:1141–1157CrossRefPubMedGoogle Scholar

Copyright information

© The Author(s) 2019

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  1. 1.Division of Digestive Surgery and UrologyTurku University HospitalTurkuFinland
  2. 2.University of TurkuTurkuFinland
  3. 3.Department of RadiologyVaasa Central HospitalVaasaFinland
  4. 4.Department of RadiologySeinäjoki Central HospitalSeinäjokiFinland
  5. 5.Department of SurgeryVaasa Central HospitalVaasaFinland
  6. 6.Department of SurgerySeinäjoki Central HospitalSeinäjokiFinland

Personalised recommendations