|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2020 | Volume
: 18
| Issue : 3 | Page : 317-324 |
|
Accuracy of MRI vs ultrasound in the diagnosis of placental adhesive disorder
Hoda A Kareem Ahmed1, Ahmed Abd-Elrady Ahmed1, Khaled Abdallah Ahmed2, Abdelgawad I Sakr1
1 Department of Radiodiagnosis, Faculty of Medicine, Al-Azhar University (Assiut Branch), Assiut, Egypt
2 Department of Obstetrics and Gynecology, Faculty of Medicine, Al-Azhar University (Assiut Branch), Assiut, Egypt
| Date of Submission | 16-Apr-2020 |
| Date of Decision | 13-May-2020 |
| Date of Acceptance | 02-Jun-2020 |
| Date of Web Publication | 30-Oct-2020 |
Correspondence Address:
Abdelgawad I Sakr
Resident of Radiodiagnosis, Assiut, 71524
Egypt
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/AZMJ.AZMJ_67_20
Background Placental adhesive disorder (PAD) is a fatal condition the rate of which has been rising progressively over the previous 50 years. PAD is the main reason for maternal morbidity and mortality and is now the most common cause of emergent postnatal hysterectomy. When transabdominal ultrasound (US) cannot conclusively exclude PAD as a diagnosis, the following imaging technique used is color Doppler ultrasonography. The gold standard investigation in high-risk patients is MRI in antenatal diagnosis.
Aim The purpose of the study was to judge the accuracy of US and color Doppler vs MRI in the prenatal diagnosis of PAD.
Patients and methods A prospective study included 50 patients aged from 20 to 40 years for high risk of developing or suspected to have PAD. The patients were evaluated by US and MRI and diagnosis was matched to postnatal outcomes.
Results A total of 50 gravid women, the age of whom ranged from 20 to 40 years were included in the study. At gestational age that ranged from 20 to 39 weeks all gravid women were examined using MRI and US. The sensitivity of the US in the evaluation of PAD was 100%, the specificity 78.95%, positive predictive value 60%, negative predictive value 100%, and the overall accuracy was about 84%. The sensitivity of MRI in the evaluation of PAD was 100%, the specificity 89.47%, positive predictive value 75%, negative predictive value 100%, and the overall accuracy was about 92%. The P value for both US and MRI diagnoses of PAD was less than 0.001.
Conclusion The US remains the primary imaging modality for assessment of the placenta. MRI has several exclusive features that make it superior in the evaluation of the placenta. Corresponding to the study, the most accurate US signs are the presence of lacunae and an unusual color Doppler imaging pattern. Corresponding to the study, the most reliable MRI signs are placental bands, heterogeneous placenta, and uterine bulging.
Keywords: placenta accreta, placenta previa, placental adhesive disorder, prenatal diagnosis
How to cite this article:
Kareem Ahmed HA, Ahmed AA, Ahmed KA, Sakr AI. Accuracy of MRI vs ultrasound in the diagnosis of placental adhesive disorder. Al-Azhar Assiut Med J 2020;18:317-24 |
How to cite this URL:
Kareem Ahmed HA, Ahmed AA, Ahmed KA, Sakr AI. Accuracy of MRI vs ultrasound in the diagnosis of placental adhesive disorder. Al-Azhar Assiut Med J [serial online] 2020 [cited 2023 Mar 26];18:317-24. Available from: http://www.azmj.eg.net/text.asp?2020/18/3/317/299577 |
| Introduction | |  |
Placental adhesive disorder (PAD) is a fatal condition the rate of which has been rising progressively over the previous 50 years. PAD is a main reason of maternal morbidity and mortality and is now the most common cause of emergent postnatal hysterectomy [1],[2].
The placenta is repeatedly overlooked in the routine assessment of an average gestation, receiving interest only when an abnormality is noticed. Although rare, abnormalities of the placenta are important to distinguish due to the possible maternal and fetal morbidity and mortality [3].
Placenta abnormalities are one of the greatest vital modern challenges in the field of obstetrics and gynecology. Placental invasion of the placenta to the uterine wall is classified on the foundation of the deepness of invasion into: Placenta accreta which occurs when the placenta is strangely adherent to the underlying decidua. Placenta increta takes place when the placenta penetrates the myometrium, placenta percreta occurs when the placenta invades the uterine serosa or the surrounding organs. PAD refers to any grade of placental invasion [4].
Placenta accreta is the abnormal implantation of the placenta into the uterine wall, and its incidence is about 0.9% of all pregnancies. Placenta previa and prior uterine surgery including cesarean delivery are risk factors [5].
Accurate prenatal diagnosis of PAD permits optimum management, guaranteeing availability of multidisciplinary team consisted of blood units, expert surgical team, anesthetist, and interventional radiologist. Transabdominal ultrasound (US) is the initial technique used to exclude PAD; however, practical difficulties present with this approach: Bladder distention and maternal obesity, posterior placentation, and contraction of the myometrium can lead to false positive results [6].
The diagnosis of PAD is made established on clinical history, imaging findings, and histopathological findings. When transabdominal US cannot conclusively exclude PAD as a diagnosis, the imaging technique used is color Doppler ultrasonography. The gold standard investigation in high-risk women is MRI in prenatal diagnosis but MRI is not as available as US and is more expensive [1].
| Aim | | |
The purpose of our study was to judge the accuracy of US and color Doppler vs magnetic resonance imaging in prenatal diagnosis of PAD.
| Patients and methods | | |
This prospective study included 50 female patients with high risk factor or suspected to have PAD. Their age ranged from 20 to 40 years. This study is conducted at Al-Azhar University Hospital, Assiut from October 2019 to April 2020. An approval by the Faculty of Medicine Ethics Committee of Al-Azhar Hospital Assiut was obtained before start of the study, and an informed written consent was signed by every patient before being enrolled into the study.
Conventional MRI and US examinations were performed for all patients. The results of the 50 patients were compared with intraoperative findings.
Inclusion criteria
History of previous uterine surgery [uterine curettage, cesarean section (CS), myomectomy], patients who got pregnant by in-vitro fertilization, smokers or hypertensive patients, patients with uterine fibroids or Asherman syndrome, history of any placental abnormality (placenta previa or placental abruption, etc.), maternal age (≥20 years) and parity (≥1).
Exclusion criteria
Patients with unreliable LMP details nor confirmed by early US, MRI contraindications: metal prosthesis, heart pacemaker in the body, and patient’s refusal for MRI examination.
Patient preparation
Urinary bladder should be moderately filled throughout MRI evaluation.
Patients were subjected to the following:- MRI studies were conducted with Philips (Philips, Achieva 1.5 Tesla MRI Unit, Netherlands). All the cases were positioned by using the abdomen coil throughout the following sequences: T2-weighted single-shot turbo spin echo: sagittal, axial, and coronal planes at right angles to the placenta–myometrium interface or myometrium–bladder interface. One breath hold T1 weighted sequence at least is taken perpendicular to the placenta. Gadolinium contrast medium was not injected in any case.
- Sonographic evaluation was done using (Siemens, Acuson ×300 US Unit, Germany). All patients were evaluated by the high-frequency transducer (3–5 MHz). Placenta should successively be evaluated in two planes: longitudinal and transverse sections to evaluate myometrium and looking for signs of placenta invasion. Serosa–bladder interface in low-lying anterior placenta and placenta previa evaluated by sagittal section just above symphysis pubis and should be evaluated along the whole interface. The urinary bladder should be moderately distended specially in the evaluation of placenta previa to evaluate serosa–bladder interface properly. During evaluation of anteriorly located placenta we should avoid compression, hence that can result in myometrial thinning and obliteration of retroplacental clear space.
- Intraoperative findings: during the surgery, PAD is confirmed by difficulty in removal of the placenta or severe bleeding once the placenta is removed which may lead to CS hysterectomy.
Statistical analysis
Statistical analysis was done using SPSS, version 25 (IBM SPSS Inc., Chicago, Illinois, USA) for Windows10. Data were expressed as mean±SD, frequency, and percentage. P value was considered significant if less than 0.05.
| Results | | |
A total of 50 gravid women in the age range from 20 to 40 years with a mean age of 29.48 years and SD 5.144 were included. At gestational age that ranged from 20 to 39 weeks all gravid were examined using MRI and US with a mean of 33.12 weeks of gestation and SD 4.885. All patients had a history of previous CS (92%), except four (8%) patients which had a history of previous D&C. Among 50 cases, placenta was located to be previa in 18 (36%) cases, anterior in 18 (36%) cases, and posterior in 14 (28%) cases.
Twenty (40%) patients had a US diagnosis of PAD (accreta/increta), while 30 (60%) patients had no signs of abnormal placentation ([Table 1] and [Figure 1]). | Table 1 Sensitivity, specificity, PPV, NPV, accuracy, and AUC of ultrasound initial diagnosis of placental adhesive disorder
Click here to view |
 | Figure 1 Receiver operating characteristic curve showing sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and area under the curve of ultrasound initial diagnosis of placental adhesive disorder.
Click here to view |
In our study, placental lacunae with turbulent color flow was found to have the highest specificity (89.47%) followed by interruption of myometrium and retroplacental vascularity (84.21%) ([Table 2]). | Table 2 Sensitivity, specificity, PPV, NPV, and accuracy of ultrasound diagnostic signs for placental adhesive disorder
Click here to view |
Sixteen (32%) patients had MRI diagnosis of PAD (accreta/increta), while 34 (68%) patients had no signs of abnormal placentation ([Table 3] and [Figure 2]). | Table 3 Sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and area under the curve of MRI initial diagnosis of placental adhesive disorder
Click here to view |
 | Figure 2 Receiver operating characteristic curve showing sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and area under the curve of MRI initial diagnosis of placental adhesive disorder.
Click here to view |
In our study, uterine bulge was found to have the highest specificity (100%), followed by heterogeneous placental intensity and dark intraplacental T2WI bands (97.37%). And in our study, interruption of the myometrium was found to have the highest sensitivity (100%) ([Table 4]). | Table 4 Sensitivity, specificity, PPV, NPV, and accuracy of MRI diagnostic signs for placental adhesive disorder
Click here to view |
Twelve cases confirmed to have placental invasion by intraoperative findings, eight (16%) cases had placenta accreta, and four (8%) cases had placenta increta.
| Discussion | | |
Imaging of the placenta have a deep effect on patient management because of the morbidity and mortality associated with different placental disorders [7].
The chief risk for PAD is prior cesarean delivery and the existence of placenta previa. The risk of placenta accreta is 24% in women with placenta previa and one previous cesarean delivery and 67% in women with placenta previa and three or more previous cesarean deliveries [2]. Other risk factors include high maternal age and a history of prior uterine surgery [2]. The surgical history of the patient is consequently very appreciated in categorizing women at highest risk for accreta [8]. In our study we found that a combination between prior CS and placenta previa increases the incidence of placenta accreta rather than other risk factors.
US and MRI are used for prenatal diagnosis of abnormal placentation, although US remains the initial investigation because it is relatively available and of low cost [9]. In our study, the US has 100% sensitivity, 78.95% specificity, 60% positive predictive value (PPV), and 100% negative predictive value (NPV). This was supported by Riteau and colleagues [10],[11],[12] who found that US has 100–89.5–93% sensitivity, 37.5–91–71% specificity, 72.2–68–74% PPV, and 100–97.6–92% NPV, respectively. However, Warshak et al. [13] and Mansour and Elkhyat [14] have found that US has 76.92–86.6% sensitivity, 96.13–80% specificity, 65.21–89% PPV, and 97.78–76.5% NPV, respectively. In our study, we found that the presence of placental lacunae with turbulent color flow ([Figure 3]), interruption of myometrium ([Figure 4]), and retroplacental vascularity ([Figure 5]) have the most specific criteria with 89.47 and 84.21% specificity, respectively, for the diagnosis of abnormal placentation. We notice that interruption of the myometrium and interruption of retroplacental vascularity have the same statistical results of 100% sensitivity, 84.21% specificity, and 88% accuracy. In the current study, loss of retroplacental clear (hypoechoic) space has the highest sensitivity (100%); however, an inadequate PPV (66.7%) for that reason detecting more false positive cases, its consistency in the diagnosis of placenta accreta is the lowest and thus it might lead to more avoidable hysterectomies. In our study, the most accurate US criterion was turbulent color flow lacunae with an NPV of 94.4%.The present study showed that retroplacental clear (hypoechoic) space between the uterus and the placenta is not reliable as a single diagnostic criterion, but should be conjoined with other criteria with greater PPV as the low PPV for loss of retroplacental clear (hypoechoic) space between the placenta and the uterus (35.3%). Additionally, the data recommended that the lack of loss of retroplacental clear (hypoechoic) space between the placenta and the uterus could help to exclude the diagnosis of PAD since they had the highest NPV of about 100% ([Figure 6]). This was supported by Comstock [15] Finberg and Williams [16] who highlighted that loss of retroplacental clear (hypoechoic) space between the placenta and the uterus reports for the bulk of false positives and this sign should not be used in isolation to establish the diagnosis. This suggested that if the retroplacental clear (hypoechoic) space between the placenta and the uterus was preserved. PAD was unlikely to occur with high NPV of 82 and 82.8%, respectively. Also Wong et al. [17] noticed loss of the clear space in 65% of patients without placenta accreta and in 100% of patients with it. Hence, the clear space concept is sensitive but not specific. The significance of it depends on its high NPV that its presence efficiently excludes placenta accreta [18]. On the other hand, some authors have pointed out that US could be imperfect in the assessment of the depth of invasion in patients suspected to have PAD. Others stated that US lacks the necessary resolution to detect PAD in posterior placenta or the patient’s body habitus represents a barrier for the scan [8]. | Figure 3 (a) Gary scale ultrasound image showing multiple irregular-shaped placental lacunae (moth-eaten appearance) (arrows), (b) color Doppler ultrasound of image (a) showing color flow in the placental lacunae.
Click here to view |
 | Figure 4 Gray scale ultrasound image showing interruption of the myometrium and loss of retroplacental hypoechoic (clear) space (arrow).
Click here to view |
 | Figure 5 Color Doppler ultrasound image showing interruption of retroplacental vascularity (arrow).
Click here to view |
 | Figure 6 Gray scale ultrasound image showing loss of retroplacental hypoechoic (clear) space (arrow).
Click here to view |
MRI can be of additional diagnostic value when additional characterization is necessary, especially in the situation of PAD. Several authors have recommended that MRI is most evidently indicated when US findings are indefinite or there is a posterior placenta. Others have instructed that MRI is superior in characterization areas of abnormal placentation, levels of invasion, and eventually adjust surgical management and should be used routinely [19]. In our study, MRI has 100% sensitivity, 89.47% specificity, 75% PPV, and 100% NPV. This was supported by Khalaf et al. [20] and Mansour and Elkhyat [14] who found that MRI has 100–93.3% sensitivity, 75–85% specificity, 93.9–94.4% PPV, and 100–82.3% NPV, respectively. However, Dwyer et al. [12] and Warshak et al. [13] have found that MRI has 80–88.46% sensitivity, 65–100% specificity, 67–100% PPV, and 79–82.35% NPV, respectively. The high-level sensitivity of MRI in the current study could be explained by the existence of more than one risk factors in our patients and most of them were suspicious for placenta accreta by US, also using SS-TSE sequences with controlled breath hold as in the Mansour and Elkhyat [14] study. Such feature provided better resolution images and almost no motion artifacts from maternal intestinal peristalsis, respiration and fetal motion. Also we depend on different sequences (T2WI TSE and T1WI SE) in different planes (coronal, sagittal, and axial) in evaluating the study cases.
In our study, we found that the presence of focal uterine bulge ([Figure 7]), the abnormal heterogeneous signal of the placenta ([Figure 8]), and dark intraplacental T2 bands ([Figure 9]) were the most specific criteria with 100, 97.37, and 97.37% specificity, respectively, for the diagnosis of abnormal placentation. Our results were in agreement with the Lax et al. [21] results who described the three criteria of invasive placentation on MRI: intraplacental T2 dark bands, heterogeneous placental signal intensity, and focal uterine bulging. Homogeneous placentas were overpoweringly benign, and significantly heterogeneous placentas with intraplacental T2 dark bands were accompanying invasive placentation [21]. It has been suggested that abnormal intraplacental T2 dark bands were caused by fibrin deposition. In contrast with Bour et al. [22] who found that placental heterogeneity was not considerably associated with the diagnosis of PAD. It is presently admitted that normal placenta shows more or less degree of physiological heterogeneity. In addition, Blaicher et al. [23] described the variety of normal presentation of the placenta. They found that placental heterogeneity depends on the age of gestation, while Varghese et al. [24] observed that the dark T2 intraplacental bands are also noticed in placental infarction and intervillous thrombus and referred to this as a popular pitfall in the diagnosis of PAD. | Figure 7 Turbo spin echo sagittal T2WI showing uterine bulging of heterogeneous placenta (arrows) in case of placenta previa.
Click here to view |
 | Figure 8 Turbo spin echo coronal T2WI showing heterogeneous placenta (arrow) and dark intraplacental bands (small arrow).
Click here to view |
 | Figure 9 Turbo spin echo sagittal T2WI showing interruption of myometrium–serosa interface with dark intraplacental band (arrow) in the case of placenta previa.
Click here to view |
Our results were in disagreement with the results of Derman et al. [2] who described primary MRI sign for the diagnosis of PAD focused on the detection of direct invasion of the placenta into the myometrium as established by thinning, unclearness of the myometrium, and loss of the thin T2 dark uteroplacental line ([Figure 10]) or direct visualization of placental tissue outside the uterus. These MRI signs were found to be nonspecific since the myometrium turn out to be very thin, particularly in the site of the cesarean scar and during the third trimester of pregnancy when it will be challenging to adequately delineate it on MRI. Additionally, the focal loss of the dark uteroplacental line on T2 SS-TSE images can be established in normal placentas and also it is not a sensitive sign when considered in separation. | Figure 10 Turbo spin echo coronal T2WI showing multiple interruption of myometrium at the uterine–bladder interface (arrow) in the case of placenta previa.
Click here to view |
| Conclusion | | |
Imaging of the placenta can have an extreme effect on patient management. It must be performed with the lowest hazard to both the mother and the fetus.
Therefore, US and MRI are noninvasive techniques and do not use ionizing radiation are favored. US remains the primary imaging modality for assessment of the placenta. MRI has several exclusive features that make it superior in the evaluation of the placenta: the multiplanar abilities, the range of pulse sequences and parameters that can be used, the better tissue contrast that can be acquired, and the absence of ionizing radiation.
Corresponding to the study, the most accurate US signs are presence of lacunae and an abnormal color Doppler imaging pattern. Subplacental clear space is less useful and should be used in combination with other findings as confirmation for PAD.
Corresponding to the study, the most reliable MRI sings are focal uterine bulging, heterogeneous signal intensity placenta, and dark T2 intraplacental bands. Focal interruption of the myometrium is less helpful because myometrium turn out to be very thin, particularly in the site of the cesarean scar and during the third trimester of pregnancy when it will be challenging to adequately delineate it on MRI.
MRI is a complementary diagnostic modality in women with high risk for PAD and must be considered when US is indecisive or inadequate. Awareness of the MRI technique to evaluate the placenta and experience with imaging findings in normal and invasive placentation will help the radiologist is contributing to an optimal outcome.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Baughman WC, Corteville JE, Shah RR. Placenta accreta: spectrum of US and MR imaging findings. Radiographics 2008; 28:1905–1916.  |
| 2. | Derman AY, Nikac V, Haberman S, Zelenko N, Opsha O, Flyer M. MRI of placenta accreta: a new imaging perspective. Am J Roentgenol 2011; 197:1514–1521. |
| 3. | Elsayes KM, Trout AT, Friedkin AM, Liu PS, Bude RO, Platt JF et al. Imaging of the placenta: a multimodality pictorial review. Radiographics 2009; 29:1371–1391. |
| 4. | Balayla J, Bondarenko HD. Placenta accreta and the risk of adverse maternal and neonatal outcomes. J Perinat Med 2013; 41:141–149. |
| 5. | Gielchinsky Y, Rojansky N, Fasouliotis SJ, Ezra Y. Placenta accreta − summary of 10 years: a survey of 310 cases. Placenta 2002; 23:210–214. |
| 6. | Elhawary TM, Dabees NL, Youssef MA. Diagnostic value of ultrasonography and magnetic resonance imaging in pregnant women at risk for placenta accreta. J Mater Fetal Neonat Med 2013; 26:1443–1449. |
| 7. | Masselli G, Gualdi G. MR imaging of the placenta: what a radiologist should know. Abdom Imaging 2013; 38: 573–587. |
| 8. | Lam G, Kuller J, McMahon M. Use of magnetic resonance imaging and ultrasound in the antenatal diagnosis of placenta accreta. J Soc Gynecol Investig 2002; 9:37–40. |
| 9. | El-Shazely MAY. Role of MRI in diagnosis of placenta accreta. J Am Sci 2014; 10:225–230. |
| 10. | Riteau AS, Tassin M, Chambon G, Le Vaillant C, de Laveaucoupet J, Quere MP et al. Accuracy of ultrasonography and magnetic resonance imaging in the diagnosis of placenta accreta. PLoS One 2014; 9:e94866. |
| 11. | Esakoff T, Sparks T, Kaimal A, Kim L, Feldstein V, Goldstein R et al. Diagnosis and morbidity of placenta accreta. Ultrasound Obstet Gynecol 2011; 37:324–327. |
| 12. | Dwyer BK, Belogolovkin V, Tran L, Rao A, Carroll I, Barth R et al. Prenatal diagnosis of placenta accreta: sonography or magnetic resonance imaging? J Ultrasound Med 2008; 27:1275–1281. |
| 13. | Warshak CR, Eskander R, Hull AD, Scioscia AL, Mattrey RF, Benirschke K et al. Accuracy of ultrasonography and magnetic resonance imaging in the diagnosis of placenta accreta. Obstet Gynecol 2006; 108(3 Pt 1):573–581. |
| 14. | Mansour S, Elkhyat W. Placenta previa-accreta: do we need MR imaging? Egypt J Radiol Nucl Med 2011; 42:433–442. |
| 15. | Comstock CH. The antenatal diagnosis of placental attachment disorders. Curr Opin Obstet Gynecol 2011; 23:117–122. |
| 16. | Finberg HJ, Williams JW. Placenta accreta: prospective sonographic diagnosis in patients with placenta previa and prior cesarean section. J Ultrasound Med 1992; 11:333–343. |
| 17. | Wong HS, Cheung YK, Zuccollo J, Tait J, Pringle KC. Evaluation of sonographic diagnostic criteria for placenta accreta. J Clin Ultrasound 2008; 36:551–559. |
| 18. | Cali G, Giambanco L, Puccio G, Forlani F. Morbidly adherent placenta: evaluation of ultrasound diagnostic criteria and differentiation of placenta accreta from percreta. Ultrasound Obstet Gynecol 2013; 41:406–412. |
| 19. | Masselli G, Brunelli R, Casciani E, Polettini E, Bertini L, Laghi F et al. Acute abdominal and pelvic pain in pregnancy: MR imaging as a valuable adjunct to ultrasound? Abdom Imaging 2011; 36:596–603. |
| 20. | Khalaf LMR, Zeid HA, Othman ER. Reliability of magnetic resonance imaging in diagnosis and assessment the depth of invasion of placental accreta in high risk gravid women. Clin Imaging 2019; 58:5–11. |
| 21. | Lax A, Prince MR, Mennitt KW, Schwebach JR, Budorick NE. The value of specific MRI features in the evaluation of suspected placental invasion. Magn Reson Imaging 2007; 25:87–93. |
| 22. | Bour L, Placé V, Bendavid S, Fargeaudou Y, Portal J-J, Ricbourg A et al. Suspected invasive placenta: evaluation with magnetic resonance imaging. Eur Radiol 2014; 24:3150–3160. |
| 23. | Blaicher W, Brugger PC, Mittermayer C, Schwindt J, Deutinger J, Bernaschek G et al. Magnetic resonance imaging of the normal placenta. Eur J Radiol 2006; 57:256–260. |
| 24. | Varghese B, Singh N, George RA, Gilvaz S. Magnetic resonance imaging of placenta accreta. Indian J Radiol Imaging 2013; 23:379–385. [ PUBMED] [Full text] |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10]
[Table 1], [Table 2], [Table 3], [Table 4]
|
![JOURNAL_FULL_NAME]](images/logo.gif){kind=logo}
Search Pubmed for