ABSTRACT
MR image data can provide many features or measures although any single measure is unlikely to comprehensively characterize the underlying morphology. We present a framework in which multiple measures are used in manifold learning steps to generate coordinate embeddings which are then combined to give an improved single representation of the population. An application to neonatal brain MRI data shows that the use of shape and appearance measures in particular leads to biologically plausible and consistent representations correlating well with clinical data. Orthogonality among the correlations suggests the embedding components relate to comparatively independent morphological features. The rapid changes that occur in brain shape and in MR image appearance during neonatal brain development justify the use of shape measures (obtained from a deformation metric) and appearance measures (obtained from image similarity). The benefit of combining separate embeddings is demonstrated by improved correlations with clinical data and we illustrate the potential of the proposed framework in characterizing trajectories of brain development.
Subject(s)
Algorithms , Brain/anatomy & histology , Brain/growth & development , Image Interpretation, Computer-Assisted/methods , Magnetic Resonance Imaging/methods , Pattern Recognition, Automated/methods , Prenatal Diagnosis/methods , Artificial Intelligence , Humans , Image Enhancement/methods , Infant, Newborn , Reproducibility of Results , Sensitivity and SpecificityABSTRACT
Functional MRI (fMRI) has not previously been used systematically to investigate brain function in preterm infants. We here describe statistically robust and reproducible fMRI results in this challenging subject group using a programmable somatosensory stimulus synchronized with MR image acquisition which induced well-localized positive blood oxygen level dependent (BOLD) responses contralateral to the side of the stimulation in: 11 preterm infants (median post menstrual age 33 weeks and 4 days, range 29+1 to 35+3); 6 control infants born at term gestational age; and 18 infants born preterm (median gestational age at birth 30 weeks and 5 days, range 25+4 to 36+0) but studied at term corrected gestational age. Bilateral signals were identified in 8 of the ex-preterm infants at term age. Anatomical confirmation of appropriate activations was provided with diffusion tensor imaging (DTI) based tractography which identified connecting pathways from the regions of activation through the ipsilateral corticospinal tracts and posterior limb of the internal capsule. These results demonstrate that it is possible to reliably identify positive BOLD signals in the infant brain and that fMRI techniques can also be applied in the study of preterm infants.
Subject(s)
Brain Mapping/methods , Infant, Newborn/growth & development , Infant, Premature/growth & development , Magnetic Resonance Imaging/methods , Somatosensory Cortex/anatomy & histology , Female , HumansABSTRACT
BACKGROUND: Very low birth weight (VLBW) infants (weight <1500 g) are increasingly cared for without prolonged periods of positive pressure ventilation (PPV). AIMS: To develop a system for 3.0 T magnetic resonance (MR) image acquisition from VLBW infants who are not receiving PPV, and to test the clinical stability of a consecutive cohort of such infants. DESIGN: Seventy VLBW infants whose median weight at image acquisition was 940 g (590-1490) underwent brain MR imaging with the developed care system as participants in research. Twenty infants (29%) received nasal continuous positive airway pressure (nCPAP), 28 (40%) received supplemental oxygen by nasal cannulae, and 22 (31%) breathed spontaneously in air during the MR examination. RESULTS: There were no significant adverse events. Seventy-six percent had none or transient self-correcting oxygen desaturations. Desaturations that required interruption of the scan for assessment were less common among infants receiving nCPAP (2/20) or breathing spontaneously in air (2/22), compared with those receiving nasal cannulae oxygen (13/28), p=0.003. Sixty-four (91%) infants had an axillary temperature > or =36 degrees C at completion of the scan (lowest 35.7 degrees C), There was no relationship between weight (p=0.167) or use of nCPAP (p=0.453) and axillary temperature <36 degrees C. No infant became hyperthermic. CONCLUSION: VLBW infants who do not require ventilation by endotracheal tube can be imaged successfully and safely at 3.0 T, including those receiving nCPAP from a customised system.
Subject(s)
Infant, Very Low Birth Weight , Magnetic Resonance Imaging/methods , Patient-Centered Care/methods , Birth Weight/physiology , Continuity of Patient Care , Humans , Infant, Newborn , Infant, Premature/physiology , Intensive Care, Neonatal/methods , Intubation, Intratracheal , Magnetic Resonance Imaging/adverse effects , Magnetic Resonance Imaging/instrumentation , Oxygen Inhalation Therapy/methods , Patient Care Team/organization & administration , Patient Positioning/methods , Positive-Pressure Respiration/methods , SafetyABSTRACT
OBJECTIVE: To assess the impact of introduction of the STAN monitoring system. STUDY DESIGN: Prospective observational study. SETTING: Tertiary referral labour ward, St George's Hospital, London. POPULATION: High-risk term pregnancies. METHODS: We report all consecutive cases of intrapartum monitoring using the STAN S 21 fetal heart monitor. Cases with adverse neonatal outcome were evaluated in relation to the ST waveform analysis and cardiotocography (CTG). MAIN OUTCOME MEASURES: Cord artery metabolic acidosis, neonatal encephalopathy (NNE) and reasons behind cases with poor outcome. RESULTS: Between 2002 and 2005, there were 1502 women monitored by STAN. Based on combined STAN analysis in the 1502 women, action was indicated in 358 women (23.8%), while in 1108 women (73.8%) no action was indicated. Traces were not interpretable in 36 women (2.4%). Of the 836 cases (55.7%) where cord blood gases were available, there were 23 cases (2.8%) of metabolic acidosis and 16 of these (70%) were identified by STAN. Overall, there were 14 cases of NNE monitored by STAN. Retrospective analysis of these highlights human errors, such as poor CTG interpretation, delay in taking appropriate action and not following the guidelines. CONCLUSIONS: Our experience suggests the need for more intense training on interpretation of CTG and strict adherence to guidelines.
Subject(s)
Acidosis/diagnosis , Cardiotocography/methods , Fetal Diseases/diagnosis , Heart Rate, Fetal/physiology , Pregnancy, High-Risk/physiology , Acidosis/physiopathology , Bradycardia/diagnosis , Bradycardia/physiopathology , Delivery, Obstetric/methods , Female , Fetal Blood/chemistry , Fetal Diseases/physiopathology , Hospitals, Teaching , Humans , Infant, Newborn , Pregnancy , Prospective StudiesABSTRACT
OBJECTIVE: We sought to evaluate whether serial amnioinfusions for persistent oligohydramnios can affect the perinatal and long-term outcomes in extreme cases of preterm premature rupture of membranes. STUDY DESIGN: All singleton pregnancies with preterm premature rupture of membranes at <26 weeks' gestation and lasting >4 days between January 1991 and June 1998 were included. Amniotic fluid volume was assessed as the maximum cord-free pocket with serial ultrasonographic examinations. Consenting women with persistent (>4 days) oligohydramnios (amniotic fluid =2 cm) received serial transabdominal amnioinfusions to maintain an amniotic fluid pocket >2 cm. The pregnancy, neonatal, and long-term neurologic outcomes of the cases that spontaneously maintained a median amniotic fluid pocket >2 cm (amnioinfusion-not-necessary group) were compared with those of women with oligohydramnios who underwent amnioinfusion but continued to have a median amniotic fluid pocket after preterm premature rupture of membranes =2 cm (persistent oligohydramnios group) and with those of women in whom oligohydramnios was alleviated by the procedure for at least 48 hours (successful amnioinfusion group). Statistical analysis included the Wilcoxon rank-sum test and the Fisher exact test, with a 2-tailed P <.05 considered significant. RESULTS: Among the 49 women included in the study, 13 (26.5%) did not have oligohydramnios, the neonatal survival rate was 92%, and normal fetal lung development and neurologic outcome were achieved in all survivors. The remaining 36 women had oligohydramnios, and all underwent serial amnioinfusions, which successfully restored a median amniotic fluid pocket >2 cm for >/=48 hours in 11 (30%) patients. This successful amnioinfusion group was comparable with the persistent oligohydramnios group (n = 25) in gestational age at first amnioinfusion (median, 20.2 weeks; range, 16-25.6 weeks; vs median, 20.3 weeks; range, 16.5-24.2 weeks; P =.4), number of amnioinfusions (median, 3; range, 1-9; vs median, 3; range, 1-5; P =.4), and interval between amnioinfusions (median, 6 days; range, 4-14 days; vs median, 8 days; range, 6-43 days; P =. 1). However, patients in the persistent oligohydramnios group had a significantly shorter interval to delivery, lower neonatal survival (20%), and higher rates of pulmonary hypoplasia (62%) and abnormal neurologic outcomes (60%) than the patients in the groups in which amnioinfusion was not necessary or was successful (all P =.01). CONCLUSION: Pregnancies with preterm premature rupture of membranes-related oligohydramnios at <26 weeks' gestation in which serial amnioinfusions successfully alleviate oligohydramnios have a perinatal outcome that is significantly better than the outcome in those with persistent oligohydramnios and is comparable with gestations with preterm premature rupture of membranes in which oligohydramnios never develops.
Subject(s)
Amniotic Fluid/physiology , Fetal Membranes, Premature Rupture/therapy , Gestational Age , Oligohydramnios/therapy , Adult , Amniotic Fluid/metabolism , Delivery, Obstetric , Embryonic and Fetal Development , Female , Fetal Membranes, Premature Rupture/complications , Humans , Infant Mortality , Infant, Newborn/physiology , Lung/embryology , Nervous System Diseases/etiology , Oligohydramnios/etiology , Pregnancy , Retreatment , Survival Analysis , Time FactorsABSTRACT
BACKGROUND: Shoulder dystocia (SD) is a rare obstetrical complication but linked with a high perinatal morbidity and mortality rate. SD has been associated to a series of maternal and fetal risk factors due to a multifactorial etiology. This study analyzes the incidence of SD, its morbidity and associated risk factors. METHODS: Cases of SD occurred at the St. Gerardo Hospital (Monza) between January 1992 trough December 1997 have been retrospectively reviewed. Obstetrical and feto-neonatal data regarding cases of SD were compared to data regarding all the cephalic vaginal deliveries occurred in the same period in our Center. RESULTS: A total of 14,157 cephalic vaginal deliveries were included in this study, of these 21 infants (0.15%) had SD. A significantly higher incidence of SD cases was found in fetal macrosomia, maternal diabetes, induction of labor by PGE2, use of obstetrical vacuum, length of first stage of labor > 4 hours in multiparas and > 8 hours in nulliparas, and length of second stage > 60 minutes, regardless of parity. A closed association was also observed between SD and birth trauma as brachial plexus injury and humerus fracture. CONCLUSION: The incidence of SD in our population (0.15%) is below the average reported in the literature (0.37-1.1%). The well-known risk factors, whose low positive predictive value can nevertheless modify obstetrical management, are confirmed.
Subject(s)
Dystocia/epidemiology , Algorithms , Female , Humans , Incidence , Infant, Newborn , Pregnancy , Retrospective Studies , Risk FactorsABSTRACT
OBJECTIVE: To assess the performance of the biophysical profile (BPP) and its components within 24 hours of delivery in predicting histopathologic evidence of severe acute placental inflammation in women with premature rupture of membranes (PROM) before 32 weeks' gestation. METHODS: We examined placentas from a series of consecutive, nonanomalous, live-born, singleton infants delivered before 32 weeks' gestation after PROM. In 166 cases, biophysical profiles (BPP) were done within 24 hours of birth. Histologic evidence of acute inflammation was assessed in the maternal (amnion) and fetal (chorionic and umbilical cord vessels) compartments, and scored on a severity scale of 0-4 by a single pathologist masked to clinical data. The presence and severity of acute inflammation was related to BPP results and its individual components. RESULTS: The overall prevalence of severe acute inflammation, ie, a score of 3 or 4, was 59% (98 of 166). In 30 (18%) cases it was present in the amnion, in 49 (30%) cases in chorionic or umbilical cord vessels, and in 19 (11%) cases in maternal and fetal compartments. There was no association between abnormal BPP score and presence or absence of severe acute placental inflammation (48% versus 46%, P =.7). Our study had a 90% power to detect a 0.26 difference between them. When rates of abnormal BPP scores were compared in cases with different degrees of acute inflammation in the maternal, fetal, or both compartments, no association was found. When the individual components of the BPP were analyzed in relation to site and severity of acute inflammation, no association was detected. CONCLUSION: We did not find evidence of a dose-response relationship between acute placental inflammation and BPP score or its individual components in cases of PROM with infants delivered before 32 weeks. Mediators other than infection might affect BPP in preterm PROM.
Subject(s)
Fetal Membranes, Premature Rupture/diagnosis , Placenta Diseases/diagnosis , Prenatal Diagnosis/standards , Acute Disease , Adult , Chorioamnionitis/diagnosis , Chorioamnionitis/epidemiology , Chorioamnionitis/pathology , District of Columbia/epidemiology , Female , Fetal Membranes, Premature Rupture/epidemiology , Gestational Age , Humans , Inflammation , Placenta Diseases/epidemiology , Placenta Diseases/pathology , Predictive Value of Tests , Pregnancy , Pregnancy Trimester, Third , Prevalence , Sensitivity and SpecificityABSTRACT
Postamniocentesis chorioamnionitis is usually managed with induction of labor to prevent maternal sepsis and related morbidity and mortality. We report a case of chorioamnionitis in a triplet pregnancy after midtrimester genetic amniocentesis, in which multiple antibiotic treatment (ampicillin 2 g i.v. loading dose followed by 1 g i.v. every 6 hr; clindamycin 900 mg i.v. every 8 hr; gentamicin 120 mg i.v. loading dose followed by 100 mg i.v. every 8 hrs; and erythromycin 500 mg i.v. every 6 hr) for 7 days and delivery of the presumably infected triplet A successfully reversed the clinical symptomatology, allowing prolongation of pregnancy until 26 weeks and survival of the remaining fetuses. At age 2 years, both infants are doing well and are meeting their developmental milestones. The viable outcome of this management strategy suggests that antibiotic treatment and expectancy may be an option in selected cases of postamniocentesis chorioamnionitis in multiple pregnancies.