Discrepant findings in a monoamniotic twin pregnancy affected by infantile myofibromatosis

Abstract

In 1876, Sir Francis Galton argued that twins are of value in studying heredity1. Monozygotic (MZ) twins are not always concordant despite deriving from one zygote. In the most comprehensive study2 12/315 monochorionic (MC) twin pregnancies were monoamniotic (MA), including four sets of conjoined twins and four sets with discrepant malformations, suggesting a rate of 50% concordance in MA non-conjoined twins. We report the case of a twin pregnancy diagnosed as being MCMA at 13 weeks' gestation. An echodense tumor was seen on the side of the head of one twin by two- and three-dimensional ultrasound examination. Amniocentesis revealed a normal karyotype (46,XX) and normal values of alpha-fetoprotein and acetylcholinesterase. Transabdominal ultrasound examination at 18 weeks, using a 3.5-MHz transducer and power Doppler imaging (Voluson 730 Expert, GE Medical Systems, Zipf, Austria), suggested vascularization of the tumor (Figure 1). Magnetic resonance imaging excluded connections to the ventricles or brain. At 30 weeks intestinal dilatation was diagnosed in the same twin. After Cesarean section at 32 weeks, the presence of a superficial head tumor was confirmed and laparotomy revealed that the intestine was compressed by a tumor mass, which was classified histologically as infantile myofibromatosis. The tumors were inoperable and the infant died from sepsis 12 days after birth. Cross-sections of the intestine and head tumor obtained at postmortem examination are shown in Figure 2. The co-twin developed uneventfully. Power Doppler ultrasound image at 18 weeks showing a head tumor with vascularization in only one of the monoamniotic twins. Congenital generalized fibromatosis was described in 19543, and later subclassified into a form limited to the skin, soft tissue or bone and a rare form with visceral involvement. Our case shows that both forms may occur sequentially. The term infantile myofibromatosis was recommended as the cells have features of fibroblasts and smooth muscle cells4. No case with infantile myofibromatosis has been diagnosed prenatally. Ten years ago, we observed one MC diamniotic twin with a mediastinal tumor leading to hydrops and postnatal death. Histology revealed infantile myofibromatosis with infiltrative growth into the lungs. The earliest that fatal infantile myofibromatosis has been diagnosed was in a 3-day-old neonate5 in whom, as in our case, histology showed proliferation of fibromatous spindle cells in the intestinal wall (Figure 2). Chemotherapy has been successfully performed postnatally in cases of infantile myofibromatosis6, and prenatal sampling or even chemotherapy might be an option in the future. Histological findings in the twin affected by infantile myofibromatosis. Cross-sections of the intestine: (a) the lumen is filled with tumor and infiltrative growth is seen within the tunica muscularis; the arrow indicates the remaining intestinal lumen with ulcerated surface; (b) detail of infiltrative growth within the tunica muscularis (arrowhead); and (c) immunohistochemical stain with antibodies against smooth muscle actin, the arrowhead indicating the tunica muscularis. In the skin biopsy of the head tumor (d), the same tumor-type as shown in the intestine is seen (lower half of image); (e) detail of the tumor found in the skin biopsy. Three families with a history of infantile myofibromatosis have been described, suggesting an autosomal dominant inheritance7 and an interstitial deletion of 6q has been detected in an infantile tumor8. Prenatal discordant malformations, and more specifically tumorigenesis, in MZ twins raises questions regarding the possible etiology, such as defects in the splitting process (with the ultimate example being conjoined twins), vascular compromise due to a shared placenta9 and de-novo mutations in one twin after splitting. Different activation of maternal and paternal genes and copy number variants have been highlighted as related to changes in gene expression, and any variation in the human genome that alters gene expression is now a candidate for being involved in a disease process10. B. Arabin [email protected]*, K. Hack*, L. Nooij*, P. Nikkels , * Department of Perinatology, Isala Clinics Zwolle, The Netherlands, Department for Pathology, University Medical Center Utrecht, The Netherlands