Virchows Archiv

, Volume 452, Issue 2, pp 201–207

Feasibility of percutaneous organ biopsy as part of a minimally invasive perinatal autopsy


  • A. C. G. Breeze
    • Division of Maternal–Fetal Medicine, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
  • F. A. Jessop
    • Department of Histopathology, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
  • A. L. Whitehead
    • Department of Histopathology, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
  • P. A. K. Set
    • Department of Radiology, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
  • L. Berman
    • Department of Radiology, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
    • Addenbrooke’s Hospital
  • G. A. Hackett
    • Division of Maternal–Fetal Medicine, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
    • Division of Maternal–Fetal Medicine, Addenbrooke’s HospitalCambridge University Hospitals NHS Foundation Trust
  • for the Cambridge post mortem MRI Study Group
Original Article

DOI: 10.1007/s00428-007-0548-7

Cite this article as:
Breeze, A.C.G., Jessop, F.A., Whitehead, A.L. et al. Virchows Arch (2008) 452: 201. doi:10.1007/s00428-007-0548-7


To determine the feasibility of percutaneous fetal organ biopsies in the context of a ‘minimally invasive’ perinatal autopsy after stillbirth and termination for abnormality is the aim of this study. We assessed successful biopsy rate and the proportion adequate for histological examination in 30 fetuses undergoing organ sampling before autopsy. The relationship between gestational age, body weight, death–biopsy interval, operator experience and successful biopsy rate was investigated. Significant findings from conventional block histology were compared with corresponding percutaneous biopsies. Of 210 organ biopsies attempted from seven target organs, 107 were obtained, of which 94 were adequate for pathological comment. The median delivery–autopsy interval was 4 (range 2–11) days. Adequate samples were obtained from the lung in 86% cases (95% CI 68, 96%), liver 76% (95% CI 56, 90%) and less frequently for the myocardium, kidney, adrenal, thymus and spleen. There was no relationship between biopsy success and time to biopsy, gestational age, body weight and user experience. No histological abnormalities found at autopsy were diagnosed from needle biopsies. Although targeted percutaneous biopsies appear feasible for some organs, fewer than 50% of all biopsies are adequate for histological examination. This technique cannot be considered to provide useful clinical information as part of a ‘minimally invasive’ perinatal autopsy.


FetusCongenital anomalyAutopsyPregnancy lossFetal deathOrgan biopsyFeasibility studies


Perinatal deaths due to fetal abnormality, late miscarriage and stillbirth complicate around 1% of pregnancies [16]. Even when there is an identifiable clinical cause for fetal death or pregnancy loss, the results of specialist investigations, including autopsy by an experienced paediatric/fetal pathologist, may modify or correct this [5], or modify the estimation of recurrence risk in cases of fetal abnormality [2]. Despite this information, increasing numbers of parents decline consent for an autopsy [11], and there are significant problems of recruitment and retention into the specialty of pediatric and perinatal pathology [5]. Furthermore, perinatal pathology services have become centralised in regional centres, requiring transport of the body of the fetus or neonate to another centre for autopsy [14].

In view of the declining post mortem consent rates and recent publicity regarding post mortem organ retention, studies have compared magnetic resonance imaging with conventional perinatal autopsy [1, 3, 4, 9, 10, 18]. Worldwide experience with perinatal post mortem magnetic resonance imaging (MRI) remains limited, and the technique remains largely unproven, with a number of questions remaining about its feasibility and implementation as a clinical service [8, 15].

A major concern relates to the inability of MRI alone to provide tissue for microscopic examination, despite providing excellent macroscopic information in many cases. The possibility of combining post mortem imaging with percutaneous organ biopsies has been raised [8, 15], thereby, providing a ‘less invasive autopsy’—but its feasibility has not been determined. This study (performed as part of an MRI-based minimally invasive autopsy), reports on the feasibility of obtaining tissue that is adequate for histological assessment from percutaneous organ biopsy and compares significant histopathological findings from organ block histology with the corresponding percutaneous biopsies.

Material and methods

This was part of a prospective comparative study from 2003 to 2005 of post mortem MRI combined with percutaneous needle biopsy vs conventional autopsy. Local research ethics committee approval was obtained (Cambridge REC reference 02/004). This institution has 5,000 maternities per year and a subregional fetal medicine unit, receiving referrals from a delivery population of approximately 24,000/year. The study group consisted of 30 fetuses, comprising those that were either miscarried or stillborn (15 fetuses) or pregnancies terminated because of fetal abnormalities (15). Gestational age ranged from 16–39 weeks and birth weight from 61–3,270 g.

Percutaneous organ biopsies were performed by an obstetric research fellow (ACGB) and specialist radiologist using ultrasound guidance ± surface landmarks in the first three cases; in the following 27 cases, they were performed using surface landmarks only. The target organs were the fetal lungs, heart, spleen, liver, kidneys, adrenals and thymus. Sampling was attempted in all cases irrespective of whether or not abnormality was suspected on ultrasound or MRI, using a 16- to 18-G ‘Temno’ cutting biopsy needle (Cardinal Health, IL, USA), a device commonly used for needle biopsy in radiology practice. These were performed just before autopsy. The biopsies were batch reported by one of two specialist perinatal pathologists, whichever did not subsequently perform the conventional post mortem.

Conventional post mortem examinations were performed by one of two consultant perinatal pathologists working in our hospital, according to Royal College of Pathologists guidelines [13]. The pathologists were informed of the clinical and ultrasound findings but were ‘blinded’ to post mortem MRI findings. The written report for the organ biopsies and conventional block histology were compared.


The median interval from delivery to autopsy was 4 days, and 95% of autopsies were performed within 7 days from delivery (range 2 – 11 days).

In one case, the parents requested biopsy blocks and slides to be returned to them before reporting. Therefore, of 210 biopsies attempted from 30 fetuses, 203 biopsies from 29 fetuses were analysed. One hundred and seven (53%) were confirmed histologically to be of the seven target organs. However, in only 94 of the 107 (46%) were these biopsies considered adequate for histopathological comment. The median number of confirmed target organ biopsies was 4 (range 0–6) per fetus. Other biopsies demonstrated connective tissue, further samples of other target organs (each organ only counting once per case) or unrecognisable tissue.

In two cases, six organs were sampled, in ten cases, five organs were successfully biopsied, in seven cases four, in three cases three, in two cases two, and in four cases only one. In one case, none of the attempted biopsies could be confirmed by microscopy. Of these 107 biopsies, 13 were considered inadequate for pathological comment.

The highest biopsy success (in terms of samples adequate for comment) was for the lungs and liver (86.0 and 76%, respectively) and lowest for the spleen (17%; see Table 1). No relationships between gestational age and success rate, nor birth weight and success rate were observed, nor for interval between death and biopsy. There was no obvious ‘learning curve’ for the technique. The regression lines in Fig. 1a,b,c and d illustrate this. In those cases in which ultrasound guidance was used, five biopsies were obtained adequate for histopathological comment in each case.
Fig. 1

a Relationship between fetal weight and number of successful biopsies obtained per fetus. Axis legends: y-axis Number of adequate biopsies obtained, x-axis fetal weight (g). b Relationship between gestational age and number of successful biopsies obtained per fetus. Axis legends: y-axis Number of adequate biopsies obtained, x-axis gestational age of fetus (completed weeks). c Relationship between number of adequate biopsies and interval between death and biopsy. Axis legends: y-axis Number of adequate biopsies obtained, x-axis interval between diagnosis of fetal death and biopsy. d Relationship between case number and number of adequate biopsies. Axis legends: y-axis Number of adequate biopsies obtained, x-axis case number

Table 1

Frequency of successful organ sampling (n = 29)


Observed organ sampling

Observed organ sampling rate for histopathological comment

Proportion (95% CIs)

Proportion (95% CIs)


0.90 (0.73, 0.98)

0.86 (0.68, 0.96)


0.83 (0.64, 0.94)

0.76 (0.56, 0.90)


0.52 (0.33, 0.71)

0.52 (0.33, 0.71


0.52 (0.33, 0.71)

0.34 (0.18, 0.54)


0.41 (0.24, 0.61)

0.41 (0.21, 0.61)


0.34 (0.18, 0.54)

0.17 (0.06, 0.36)


0.17 (0.06, 0.36)

0.17 (0.06, 0.36)

Table 2 shows the potentially significant findings at block histology and the corresponding findings (if available) of the percutaneous core biopsies from the same fetus. In most cases, organ histology was normal, or organs were too autolysed for comment at full post mortem. However, there was one case with significant liver block findings, five cases with significant lung block findings, one case with significant myocardial block findings, one case with significant renal block findings, two cases with adrenal block findings, one case with significant spleen findings, and three cases with significant thymus block findings. In no cases were the significant block histology findings replicated in the organ biopsies; these were all reported as either normal, inadequate or no biopsy available.
Table 2

Correlation between significant block histology findings and corresponding percutaneous biopsy


Significant block findings

Corresponding biopsy findings


Case 1—chronic inflammatory infiltrate around bronchioles extending into lung parenchyma

Case 1—no biopsy available

Case 2—focal necrosis

Case 2—apparently normal lung

Case 4—focal necrosis

Case 4—lung biopsy inadequate

Case 8—interstitial inflammatory cell infiltrate

Case 8—apparently normal lung

Case 21—poorly formed potential airspaces

Case 21—lung biopsy inadequate


Case 28—Congested liver

Case 28—normal liver


Case 4—focal necrosis

Case 4—no biopsy available


Case 16—multicystic dysplastic kidneys

Case 16—no biopsy available


Case 1—congested, haemorrhagic adrenals

Case 1—no biopsy available

Case 11—haemorrhagic adrenals

Case 11—normal adrenals


Case 11—cortical depletion

Case 11—normal

Case 23—cortical depletion

Case 23—inadequate biopsy

Case 29—large cystic Hassall’s corpuscles

Case 29—no biopsy available


Case 11—haemorrhagic spleen

Case 11—no biopsy available


The aim of this study was to determine whether samples adequate for pathological comment could be obtained by percutaneous needle sampling from fetuses after stillbirth or termination for abnormality. The two questions addressed by this study are represented by Tables 1 and 2, namely, how frequently are adequate samples obtained from fetal organs and how do percutaneous biopsy results compare with those of conventional block histology?

While for some fetal organs, such as lungs and liver, adequate samples could be obtained in the majority of cases, for others, most samples were inadequate. These difficulties reflect the relatively small size of many fetal organs, particularly in the mid-trimester and possibly also the lack of imaging guidance in this study. It is also noteworthy that where histological abnormalities were found at conventional autopsy, in no cases were these identified in the needle biopsies taken of the same organs—either because the biopsies failed, were inadequate or were reported as normal. This must be regarded as a significant weakness of organ biopsy in this study; in comparison with histology from conventional autopsy, the false negative rate of percutaneous biopsy was 100%.

In many perinatal autopsies, organ histology does not reveal histological abnormalities in most organs; a frequent finding is that many internal organs are too autolysed for pathological comment. In cases of termination for fetal abnormality, histological abnormalities may only be expected in a small number of organs, usually suspected in advance on prenatal imaging. In antepartum fetal death, histological abnormalities may be most commonly found in the liver or lungs, for example, showing evidence of fetal bacterial or viral infection [16]. In this context, adequate samples were most easily obtained from the fetal lungs and liver potentially allowing assessment of fetal organ development and maturation—and the histological investigation of perinatal infection. However, none of the biopsies in this study revealed such findings, even when, for example, block histology from the same fetus showed evidence of infection. Abnormalities in the fetal thymus (“cortical depletion”) may be seen in cases of chronic hypoxia. The thymus was successfully sampled in only 17% of biopsy attempts, and fetal spleen and adrenal were sampled with most difficulty. We must therefore be guarded about recommending percutaneous biopsy of any fetal organs after stillbirth as the likelihood of successful biopsy for some organs, and then disclosing abnormalities may be low.

The results presented are comparable with those described previously for adults, with adequate samples obtained from the liver in 92% of cases, the heart in 55%, the lung in 46% and the kidney in 34% [17]. A more recent—but smaller—study reported successful biopsy of the adult lung in 100% of 21 cases to only 9.5% for the kidney [7]. To our knowledge, the only previous reported data on needle biopsy in the context of perinatal loss was that described in one neonate after death after neonatal encephalopathy [12]; it is unclear whether imaging guidance was used in this case. The only experience of percutaneous organ biopsy in children used ultrasound guidance to obtain biopsies of organs in 19 children [6], although the data reported do not allow direct comparisons with this study, as they concentrated on agreement on cause of death.

The success rates for organ sampling that we describe provide pragmatic data on what could be expected to be achieved without highly trained individuals or imaging guidance and suggest that tissue can be obtained by percutaneous biopsy of certain major fetal organs using surface landmarks. There was no apparent learning curve relating to increasing experience. Our data suggest that we must be cautious about the interpretation of apparently adequate percutaneous biopsies, as in all cases in which there were histological abnormalities on block histology, these were not detected by biopsy. Furthermore, in many of the cases in which significant histological abnormalities were identified on conventional organ blocks, inadequate or no tissue from the organ of interest was obtained by percutaneous biopsy (Table 2). While it is possible, therefore, that imaging-directed biopsy (whether by ultrasound or MRI) might provide a higher proportion of samples that are both adequate for comment, and able to replicate the findings of block histology, this remains to be proven.

This paper details the feasibility of percutaneous organ biopsies as part of a “minimally invasive” perinatal autopsy—not its comparability to conventional autopsy. Although approximately one half of the organs were successfully sampled, even where this was the case, it did not allow an accurate histological diagnosis to be made. We detected none of the histological abnormalities that were found by conventional tissue-block histology, although the sample of abnormal cases is admittedly small. We therefore demonstrate no utility for percutaneous needle biopsy. Whether percutaneous biopsy would make any practical difference to achieving a diagnosis and counselling parents where specific pathologies were suspected remains to be demonstrated. Furthermore, the role of percutaneous biopsies as a stand alone investigation or in combination with imaging in perinatal death would appear on the basis of these findings to be limited.


Members of the Cambridge post mortem MRI study group also include: Dr Justin Cross, Professor David J Lomas, and Ms Ilse Joubert (of the Department of Radiology, University of Cambridge and Addenbrooke’s Hospital). We are indebted to the parents who agreed to take part in this study at a very difficult time.

We also acknowledge the assistance of Andrew F Dean (Department of Histopathology, Addenbrooke’s Hospital), and Dr Nicholas Coleman (Hutchison-MRC Centre, Cambridge), and the paediatric pathology service staff of Addenbrooke’s Hospital: Nicola Wood, Gillian Kenyon, Simon Brown, and Michelle Macer. We also wish to thank the midwives and medical staff of the Rosie Hospital, Cambridge for their support.

This study was supported by a grant from the Trustees of the Addenbrooke’s Charities, and ACGB’s salary was supported by Cambridge Fetal Care.

Conflict of interest statement

The authors have no conflicts of interest to declare.

Copyright information

© Springer-Verlag 2007