The Twin Peak" Sign: 11
Reliable Evidence of Dichorionic Twinning Harris J. Finberg, MD
In twin pregnancies with a single placental zone, the presence of a triangular projection of placental tissue beyond the chorionic surface, extending between the layers of the intertwin membrane, has provided reliable evidence that there are two fused placentas (dichorionic, diamniotic) rather than a single shared placenta (monochorionic,. diamniotic). This observation, dubbed the Mtwin peak~ sign, was identified in 15 twin pregnancies, all proven to be dichorionic at birth, and in five triplet pregnanCies, all proven to be trichorionic. This finding is produced by proliferating chorionic villi
I
nterest in the ability to differentiate reliably and accurately between dichorionic and monocho~ rionic twin pregnancies sonographically is due to the substantially higher risks of morbidity and mortality in multiple gestations with a single shared placenta. When two geographically distinct placentas are present, the pregnancy is dichorionic. When there is a single placental region, however, the pregnancy may have either one placenta shared by both twins or two separate placentas that have grown to be contiguous and have become fused. This report describes an observation, dubbed the •twin peak ~ sign, which, when present, confirms that the twin pregnancy is dichorionic. This finding may also be applied in triplet pregnancies. The anatomic basis for this sign is elucidated, explaining its reliability.
Rt:cl!'iwd April 3, 1992, from Phoenix Pl!'rinatal Associates, Phoe· nix, Arizona. Rl!'vist:d manuscript acct:pted for publication Jun~ 23, 1992. Addrt.'ss corrt.'spondt.'nCt.' and reprint requests to Harris ). Finbl!'rg, MD, Director, Diagnostic Ultrasound. Phoenix Perinatal Associates, 1300 Nmth 12th Street, /1320, Phoenix., AZ 85006.
growing into the potential space between the two layers of chorion in the intertwin membrane. The single chorion of a monoplacental twin pregnancy serves as an intact barrier, preventing villi from growing between the two amniotic layers. A decision sequence incorporating the use of the twin peak sign is described for determining the type of twinning that has occurred. KEY woRos: Pregnancy; Multiple gestation; Dichorionic twinning; Placenta; Obstetrical ultrasound; Twin peak sign.
METHODS Twin and triplet pregnancies in second and third trimesters were scanned by the author, using a variety of commercially available ultrasound systems with transducer frequencies ranging from 3.5 MHz to 7.5 MHz. The observation sought was the presence of a triangular projection of tissue with the same echogenicity as the placenta extending beyond the chorionic surface of the placenta, insinuated between the layers of the intertwin membrane, wider at the chorionic surface and tapering to a point at some distance inward from that surface (Fig. 1). All cases that showed this finding, referred to as the •twin peak• sign in this report, whether over the entire length of the placentamembrane junction or over a short segment only, were included for eventual postnatal confirmation of chorionicity through the histologic evaluation of the intertwin membrane. Birth of male-female twin pairs was used as proof in two cases early in the series for which histologic reports were not available. Multiple gestations in which this finding was not detected along the visualized placenta-membrane june· tion (Fig. 2) or in which the placental surface was obscured by overlying fetuses were not included for
© 1992 by the American Institute of Ultrasound in Medidne • J Ultrasound Med 11 :571- 577, 1992 • 0278-4297/ 92/ $3:50
572
THE TWIN PEAK SIGN
A
J Ultrasound Med 11:571- 577, 1992
Figure 2 No twin peak sign is seen at the placental-septum junction. The chorionicity of this twin pregnancy is., therefore, indeterminate by the twin peak criterion. Such cases were not included in the data collection for this series. (The thin and wispy nature of the membrane strongly suggests it to be monochorionic twinning.)
these early stages, but the observation did not appear to add discriminatory ability beyond that by identifying the sacs as separate.
RESULTS
B Figure 1 The twin peak sign. An extension of placental tissue is seen widening the intertwin septum at its pojnt of contact with the chorionic surface of the placenta.. The peak is triangular in cross section, widest at the placental surface and tapering to a point at some distance inward along the membrane. A, 15 week dkhorionic twins. B, 27 week dizygotic male-female twins.
follow~up. As a consequence, the frequency of occurrence of this sign in dichorionic pregnancy and the likelihood that a pregnancy without this observation will be monochorionic are not addressed quantitatively in this report. Scans of twin and triplet pregnancies done prior to a gestational age of 12 weeks (menstrual dating) were excluded from the study group because the individual gestational sacs generally were stiU recognizable. Presence of decidual tissue extending between gestational sacs, equivalent to the finding reported in this study, was seen in many of the multichorionic pregnancies at
The twin peak sJgn was identified Jn 15 twin pregnan· cies and five triplet pregnancies on scans performed between 14 and 35 weeks of gestati.on age. The first sonogram obtained in the second or third trimester that showed the chorionic peak finding occurred at 13 to 17 weeks in seven patients, at 18 to 22 weeks in six patients, at 23 to 27 weeks in two patients, and at 28 to 35 weeks in five pati.ents. Most multiple pregnancies referred to our practice are first seen in second trimes· ter. Peaks remained detectable on later scans in many cases, although crowding by fetal parts obscured the posterior placentas progressively more completely with advancing gestational age. The size of the peaks varied (Fig. 3), and the peaks were most often seen only segmentally rather than continuously along the interface of the intertwin mem· brane with the placenta. In each of the triplet pregnancies, the twin peak sign was identified in at least two separate membrane interfaces (e.g., along the mem~ brane between fetuses A and B and also along the membrane between fetuses B and C) (Fig. 4), except for one case with one separate and one fused dicho~ rionic placenta. All 15 twin pregnancies were proven to be dicho· rionic, 13 by postpartum histologic inspection of the intertwin membrane. Two cases early in the series did
J Ultrasound Med 11:571-577, 1992
Figure 3 The smallest twin peak identified in this series
(arrow). The patient was obese and at 35 weeks of gestation.
FINBERG
573
A
A previous scan was interpreted, by a different observer, as being monochorionic based on the relative thinness of the membrane and the difficulty in visualizing it. The pregnancy was confirmed to be dichorionic at delivery by membrane histology.
B
Figure 4 Fused trichorionic triplet placenta at 16 weeks, with twin peaks seen in septum between triplets A and B and between triplets B and C.
not have histologic results available, but both were male-female twin pairs, indicating these pregnancies to be dizygotic and hence dichorionic. All five triplet pregnancies had postpartum membrane histologic findings that confirmed the trichorionic status. Several variations were noted in the twin peak find ings in this series. Segmental presence of a hyperechoic or hypoechoic linear interface within the twin peak, aligned roughly with the intertwin septum, was seen in four cases (Fig. 5). A venous lake was present at the base of the twin peak in one. The placental origin of
Figure 5 Linear interfaces detected within the substance of the twin peak, aligned with the orientation of the twin membrane. These ·boundaries• were found in some but not all regions of the peak in the four cases in which this phenomenon was noted. No true membrane separates the chorionic villi from the two fused placental moieties. It is speculated that the alignment or organization of the villi from each portion of the placenta occasionally creates a large enough zone of planar interface to be imaged by ultrasonog·· raphy. A, Hyperechoic linear interface within the peak (ar· rows). B, Hypoechoic linear zone within the peak (arrowlleads).
the umbilical cord of one twin was noted along the surface of the twin peak in one case, representing a marginal cord insertion relative to that twin's placenta. There were no unusual events in any of these preg· nancies that might suggest any clinical importance to these variations. All twins and triplets in this series
574
JUltrasound Med 11:571- 577, 1992
THE TWIN PEAK SIGN
were liveborn and healthy without unusual prematurity and without pregnancy complications that might be attributable to the twin peak sign.
DISCUSSION Presence of the twin peak sign was associated with dichorionic placentation in the twins and with trichorionic placentation in the triplet pregnancies in every case in this series. The uniform success of the sign is explained by the anatomic basis for the finding. The twin pei\k sign represents the extension of placental villi into the potential interchorionic space of the twin membrane. Th•s space exists only in dichorionic pregnancy and is produced by reflection of each chorion away from its placenta at the place where it encounters the chorion and placenta of the co-twin. The villi of each placenta are in direct contiguity with each other, leading to fusion of the two placental masses, but with extremely rare exceptions, no vascular communication occurs between them. As the villi from either or both placental moieties continue to proliferate, they can grow out into the interchorionic space of the twin membrane (Fig. 6A). The presence of placental villi in the interchorionic space could alternatively be due to incomplete regres~ sion of decidua basalis from the site where the intertwin membrane first develops, but the size of the twin peaks observed would imply some subsequent villous growth, even if that is in situ. The constant triangular shape, wider at the chorionic surface of the placenta and tapering inward, also is more consistent with growth into the interchorionic space from the placenta rather than persistence of villous islands, which might occasionally appear globular with narrow zones of connection to the main portion of the placenta, like a succenturiate lobe. The twin peak sign cannot occur in monochorionic placentation because the single continuous chorion serves as an impenetrable barrier to the growth of the placental villi. These villi are thus excluded from extension into the potential interamniotic space of the monochorionic, diamniotic twin membrane (Fig. 68). The presence of placental villi in the interchorionic space of the dichorionic, diamniotic twin membrane can easily be identified by gross inspection of the placenta at delivery (Fig. 7), It is important to note that in the gross specimen (as in the illustrated case) and on sonograms during pregnancy, the interchorionic extension of placental villi usually appears focally scattered along the placental surface origin of the intertwin membrane, rather than occurring as a continuous ridge. In seeking the twin peak sign by ultrasonography, it is therefore imperative that the zone of intersection of the membrane with the placenta be carefully scruti~ nized over its entire length (or over as much as can be
A
A
1
B Figure 6 Chorionic and amniotic relationships explain the occurrence of the twin peak phenomenon only in dichorionic twin pregnancies. A, In a dichorionic pregnancy with fused placentas, both the amnions (A) and the chorions (C) reflect away from the placental surface at the point of origin of the septum. This creates a potential space in direct continuity with the chorionic villi and into which they can extend. B, In a monochorionic twin pregnancy, the septum is formed by reflection of the two amnions away from the placenta. There is a continuous single chorion, which provides an intact barrier, preventing extension of placental. villi into the potential interamniotic space.
detected, if portions are obscured by overlying fetal parts). It is important to recognize that absence of the twin peak sign does not imply that the pregnancy is mono· chorionic. When no chorionic peak is identified, the chorionid ty of the pregnancy remains indeterminate by this criterion. This is analogous to the decision pattern of identifying a twin pregnancy with two
J Ultrasound Med 11:571 - 577, 1992
Figure 7 Fused dichorionic twin placenta specimen after delivery. The segmental presence of placental villi are easily seen within the twin septum. This is the anatomic basis of the sonographic twin peak sign. (Photograph courtesy of Kurt Benirschke, MD, Professor of Pathology, University of California, San Diego.)
distinct placentas as dichorionic while the chorionicity of twins with a single placental zone cannot be distinguished, except by other criteria. As noted in the methodology, the data collected did not permit analysis for the frequency of this sign in dichorionic, fused-placenta pregnancies. Nonetheless, by comparison with observations of other signs, including membrane thickness and discordant sex twin pairs, it appeared that twin peaks are detectable in the majority of such cases when a diligent search for their presence is made. A possible pitfall in using chorionic peaks to assign chorionicity to triplet pregnancies should be noted. If three placentas have fused as adjacent bars or strips, identification of one peak between triplets A and B and a second peak between B and C is sufficient for confirming trichorionic pregnancy (Fig. SA). If, instead, the placentas are arranged radially, like pie wedges, peaks occurring between triplets A and B and between B and C would prove only that triplet B has a separate placenta from A and C, whereas triplets 'A and C could still share a monochorionic placenta (Fig. SB). The geometry of fused triplet placentas may be complex and not fully assessable sonographically. A second potential pitfall could occur if a uterine synechia was present, indenting the gestational sac of a monochorionic twin pregnancy in the region of the placenta. In over half of the cases in which a synechia is seen in pregnancy the placenta extends to, wraps around, or grows out along the adhesion. 1 This can create a peak-like elevation of the placenta which, in a twin pregnancy, could simulate the twin peak sign. It would, however, have no relationship to the intertwin membrane, except by chance proximity.
FINBERG
575
A
c
B
A B Figure 8 Variations in the geometry of fused triplet placentas. A, The three placental components are arranged as parallel bars or strips. Presence of twin peaks both between triplets A and B and between triplets B and C confirms the pregnancy to be trichorionic. 8, The three placental components are arranged radially. In this geometric arrangement, twin peaks between triplets A and B and between triplets B and C confirm only that B has a separate placenta from A and C. No statement about the placental interface between A and C can be made. These two may share a monochorionic placenta or have fused dichorionic placentas.
The twin peak phenomenon can occur at the margin of a placenta in a twin pregnancy with two distinct placentas if the intertwin membrane happens to originate at this site. (One such case, not included in this series, was seen.) Such an observation is not generally
576
THE TWIN PEAK SIGN
of diagnostic import. If there were potentia] concern that two separate placental zones represented a monochorionic twin placenta with a large succenturiate lobe, as in a case of significant size discordance of like· sex twins, presence of a twin peak would effectively exclude that possibility. The widespread interest in determining the type of placental development in twin pregnancies is predi+ cated on the substantially greater risk of shared-placenta twinning. Benirschke and Kim in 1973 2 reported a perinatal mortality of 26% in diamniotic, monochorionic twins as compared to 9% in diamniotic, dichorionic twins. (The mortality rate in the fortunately uncommon monoamniotic, monochorionic twin pregnancies was SO%.) Beginning with Barss and colleagues in 1985, recent efforts at distinguishing monochorionic from dichorionic twinning when separate placentas are not seen have focused on categorizing the twin septum as thin (monochorionic) or thick (dichorionic). 3 Hertzberg and coworkers found that accuracy in using septum thick· ness diminished with advancing pregnancy stage! Townsend and associates reported an 83% predictive value for both monochorionic and dichorionic twin pregnancies using membrane thickness criteria:" D' AJton and Dudley6 correctly determined chorionicity in 68 of 69 twin pregnancies by counting the number of layers in the intertwin membrane as seen on the son· agraphic image. Two layers indicated a monochorionic pregnancy, and three or four layers, a dichorionic set of twins. 6 In our own laboratory, we have not been able to verify this observation. The twin peak sign described in this report has received little if any attention in the ultrasound literature. A brief report by Bessis and Papiernik in 198t1 mentioned a '"lambda sign" as present in 20 of 24 dichorionic pregnancies and in none of six monochorionic twin pregnancies. They do not describe the find ing but have two labeled illustrations, one at 10 weeks and one at 32 weeks, that show a finding compatible with that described in the present report. Determination of the Type of Twinning To determine the type of twinning that has occurred, the following decision cascade is suggested. 1. At 10 weeks or Jess, count the number of gesta• tional sacs. Presence of two sacs, each with a live fetus, indicates dichorionic twinning. A single sac within which two live fetuses are present is monochorionic. Beyond 10 weeks the following order of criteria can be used. 2. Determine the gender of the twins. A male-female pair proves a dizygotic and, therefore, a dichorionic pregnancy. Uke sex or unidentified gender twins are of indeterminate chorionicity.
J Ultrasound Med 11:571- 577, 1992
3. Determine the number of separate placentas. Two distinct placentas indicate a dichorionic pregnancy. A single placental region is indeterminate, occurring in either a monochorionic or a fused dichorionic pregnancy. 4. With one placental zone seen, search for a twin peak sign. If it is present, the pregnancy has a dichorionic fused placenta. If none is detected, the chorionidly remains undetermined. 5. If the previous criteria have not identified the chorionic status, evaluate the intertwin membrane. A thick membrane indicates a high probability of dicho· rionic twinning, whereas a thin and wispy membrane implies a simi1arly high probability of a monochorionic twin pair. 6. If no membrane is detected, a monoamniotic, monochorionic twin pregnancy must be considered. The septum (usually of a diamniotic, monochorionic set of twins) may occasionally be present but undetected because of technical factors 8 or as a result of oligohydramnios within the sac of one twin so that the septum is so closely applied to this twin that it is not visualized . Such a twin is generally held tightly against the uterine wall, and is referred to as a •stuck twin ." This phenomenon can be demonstrated by having the mother roll so that this twin is nondependent and observing that the twin does not shift downward with gravity. Presence of entangled umbilical cords can con• firm monoamniotic twinning. 9 If these findings are inconclusive, single sac amniography with computed tomographic detection of intestinal contrast material in both twins can be used to prove that the pregnancy is monoamniotic, monochorionic. 111
REFERENCES 1. Finberg HJ: Uterine synechiae in pregnancy: Expanded criteria for recognition and review of clinical significance in 28 cases. J Ultrasound Med 10:547, 1991 2. Benirschke K, Kim CK: Multiple pregnancy (first of two parts). N Eng! J Med 288:1276, 1973 3. Barss VA, Benacerraf BR, Frigoletto FD: Ultrasonographic determination of chorion type in twin gestation. Obstet Gynecol66:779, 1985 4. Hertzberg BS, Kurtz AB, Choi HY, et al: Significance of membrane thickness in the sonographic evaluation of twin gestations. AJR 148:151, 1987 5. Townsend RR, Simpson GF, Filly RA: Membrane thickness in ultrasound prediction of chorionicity of twin gestations. J Ultrasound Med 7:327, 1988 6. D'Alton ME, Dudley DK; The ultrasonographic prediction of chorionicity in twin gestation. Am J Obstet Gy• necol160:557, 1989
J Ultrasound Med 11:571-577, 1992
7. Bessis R, Papiemik E: Echographic imagery of amniotic membranes in twin pregnancies. In Twin Research 3: Twin Biology and Multiple Pregnancy. New York, Alan R. Liss, 1981, p 183 8. Mahony 85, Filly RA, Callen PW: Amnionicity and chorionicity in twin pregnancies: Prediction using ultrasound. Radiology 155:205, 1985
FINBERG
577
9. Nyberg DA, Filly RA, Golbus MS, et al: Entangled urn• bilical cords: A sign of monoamniotic twins. j Ultrasound Med 3:29, 1984
10. Finberg HJ, Clewell WH: Definitive prenatal diagnosis of monoamniotic twins: Swallowed amniotic contrast agent detected in both twins on sonographically selected CT images. J Ultrasound Med 10:5123, 1991
Reliable Evidence of Dichorionic Twinning Harris J. Finberg, MD
In twin pregnancies with a single placental zone, the presence of a triangular projection of placental tissue beyond the chorionic surface, extending between the layers of the intertwin membrane, has provided reliable evidence that there are two fused placentas (dichorionic, diamniotic) rather than a single shared placenta (monochorionic,. diamniotic). This observation, dubbed the Mtwin peak~ sign, was identified in 15 twin pregnancies, all proven to be dichorionic at birth, and in five triplet pregnanCies, all proven to be trichorionic. This finding is produced by proliferating chorionic villi
I
nterest in the ability to differentiate reliably and accurately between dichorionic and monocho~ rionic twin pregnancies sonographically is due to the substantially higher risks of morbidity and mortality in multiple gestations with a single shared placenta. When two geographically distinct placentas are present, the pregnancy is dichorionic. When there is a single placental region, however, the pregnancy may have either one placenta shared by both twins or two separate placentas that have grown to be contiguous and have become fused. This report describes an observation, dubbed the •twin peak ~ sign, which, when present, confirms that the twin pregnancy is dichorionic. This finding may also be applied in triplet pregnancies. The anatomic basis for this sign is elucidated, explaining its reliability.
Rt:cl!'iwd April 3, 1992, from Phoenix Pl!'rinatal Associates, Phoe· nix, Arizona. Rl!'vist:d manuscript acct:pted for publication Jun~ 23, 1992. Addrt.'ss corrt.'spondt.'nCt.' and reprint requests to Harris ). Finbl!'rg, MD, Director, Diagnostic Ultrasound. Phoenix Perinatal Associates, 1300 Nmth 12th Street, /1320, Phoenix., AZ 85006.
growing into the potential space between the two layers of chorion in the intertwin membrane. The single chorion of a monoplacental twin pregnancy serves as an intact barrier, preventing villi from growing between the two amniotic layers. A decision sequence incorporating the use of the twin peak sign is described for determining the type of twinning that has occurred. KEY woRos: Pregnancy; Multiple gestation; Dichorionic twinning; Placenta; Obstetrical ultrasound; Twin peak sign.
METHODS Twin and triplet pregnancies in second and third trimesters were scanned by the author, using a variety of commercially available ultrasound systems with transducer frequencies ranging from 3.5 MHz to 7.5 MHz. The observation sought was the presence of a triangular projection of tissue with the same echogenicity as the placenta extending beyond the chorionic surface of the placenta, insinuated between the layers of the intertwin membrane, wider at the chorionic surface and tapering to a point at some distance inward from that surface (Fig. 1). All cases that showed this finding, referred to as the •twin peak• sign in this report, whether over the entire length of the placentamembrane junction or over a short segment only, were included for eventual postnatal confirmation of chorionicity through the histologic evaluation of the intertwin membrane. Birth of male-female twin pairs was used as proof in two cases early in the series for which histologic reports were not available. Multiple gestations in which this finding was not detected along the visualized placenta-membrane june· tion (Fig. 2) or in which the placental surface was obscured by overlying fetuses were not included for
© 1992 by the American Institute of Ultrasound in Medidne • J Ultrasound Med 11 :571- 577, 1992 • 0278-4297/ 92/ $3:50
572
THE TWIN PEAK SIGN
A
J Ultrasound Med 11:571- 577, 1992
Figure 2 No twin peak sign is seen at the placental-septum junction. The chorionicity of this twin pregnancy is., therefore, indeterminate by the twin peak criterion. Such cases were not included in the data collection for this series. (The thin and wispy nature of the membrane strongly suggests it to be monochorionic twinning.)
these early stages, but the observation did not appear to add discriminatory ability beyond that by identifying the sacs as separate.
RESULTS
B Figure 1 The twin peak sign. An extension of placental tissue is seen widening the intertwin septum at its pojnt of contact with the chorionic surface of the placenta.. The peak is triangular in cross section, widest at the placental surface and tapering to a point at some distance inward along the membrane. A, 15 week dkhorionic twins. B, 27 week dizygotic male-female twins.
follow~up. As a consequence, the frequency of occurrence of this sign in dichorionic pregnancy and the likelihood that a pregnancy without this observation will be monochorionic are not addressed quantitatively in this report. Scans of twin and triplet pregnancies done prior to a gestational age of 12 weeks (menstrual dating) were excluded from the study group because the individual gestational sacs generally were stiU recognizable. Presence of decidual tissue extending between gestational sacs, equivalent to the finding reported in this study, was seen in many of the multichorionic pregnancies at
The twin peak sJgn was identified Jn 15 twin pregnan· cies and five triplet pregnancies on scans performed between 14 and 35 weeks of gestati.on age. The first sonogram obtained in the second or third trimester that showed the chorionic peak finding occurred at 13 to 17 weeks in seven patients, at 18 to 22 weeks in six patients, at 23 to 27 weeks in two patients, and at 28 to 35 weeks in five pati.ents. Most multiple pregnancies referred to our practice are first seen in second trimes· ter. Peaks remained detectable on later scans in many cases, although crowding by fetal parts obscured the posterior placentas progressively more completely with advancing gestational age. The size of the peaks varied (Fig. 3), and the peaks were most often seen only segmentally rather than continuously along the interface of the intertwin mem· brane with the placenta. In each of the triplet pregnancies, the twin peak sign was identified in at least two separate membrane interfaces (e.g., along the mem~ brane between fetuses A and B and also along the membrane between fetuses B and C) (Fig. 4), except for one case with one separate and one fused dicho~ rionic placenta. All 15 twin pregnancies were proven to be dicho· rionic, 13 by postpartum histologic inspection of the intertwin membrane. Two cases early in the series did
J Ultrasound Med 11:571-577, 1992
Figure 3 The smallest twin peak identified in this series
(arrow). The patient was obese and at 35 weeks of gestation.
FINBERG
573
A
A previous scan was interpreted, by a different observer, as being monochorionic based on the relative thinness of the membrane and the difficulty in visualizing it. The pregnancy was confirmed to be dichorionic at delivery by membrane histology.
B
Figure 4 Fused trichorionic triplet placenta at 16 weeks, with twin peaks seen in septum between triplets A and B and between triplets B and C.
not have histologic results available, but both were male-female twin pairs, indicating these pregnancies to be dizygotic and hence dichorionic. All five triplet pregnancies had postpartum membrane histologic findings that confirmed the trichorionic status. Several variations were noted in the twin peak find ings in this series. Segmental presence of a hyperechoic or hypoechoic linear interface within the twin peak, aligned roughly with the intertwin septum, was seen in four cases (Fig. 5). A venous lake was present at the base of the twin peak in one. The placental origin of
Figure 5 Linear interfaces detected within the substance of the twin peak, aligned with the orientation of the twin membrane. These ·boundaries• were found in some but not all regions of the peak in the four cases in which this phenomenon was noted. No true membrane separates the chorionic villi from the two fused placental moieties. It is speculated that the alignment or organization of the villi from each portion of the placenta occasionally creates a large enough zone of planar interface to be imaged by ultrasonog·· raphy. A, Hyperechoic linear interface within the peak (ar· rows). B, Hypoechoic linear zone within the peak (arrowlleads).
the umbilical cord of one twin was noted along the surface of the twin peak in one case, representing a marginal cord insertion relative to that twin's placenta. There were no unusual events in any of these preg· nancies that might suggest any clinical importance to these variations. All twins and triplets in this series
574
JUltrasound Med 11:571- 577, 1992
THE TWIN PEAK SIGN
were liveborn and healthy without unusual prematurity and without pregnancy complications that might be attributable to the twin peak sign.
DISCUSSION Presence of the twin peak sign was associated with dichorionic placentation in the twins and with trichorionic placentation in the triplet pregnancies in every case in this series. The uniform success of the sign is explained by the anatomic basis for the finding. The twin pei\k sign represents the extension of placental villi into the potential interchorionic space of the twin membrane. Th•s space exists only in dichorionic pregnancy and is produced by reflection of each chorion away from its placenta at the place where it encounters the chorion and placenta of the co-twin. The villi of each placenta are in direct contiguity with each other, leading to fusion of the two placental masses, but with extremely rare exceptions, no vascular communication occurs between them. As the villi from either or both placental moieties continue to proliferate, they can grow out into the interchorionic space of the twin membrane (Fig. 6A). The presence of placental villi in the interchorionic space could alternatively be due to incomplete regres~ sion of decidua basalis from the site where the intertwin membrane first develops, but the size of the twin peaks observed would imply some subsequent villous growth, even if that is in situ. The constant triangular shape, wider at the chorionic surface of the placenta and tapering inward, also is more consistent with growth into the interchorionic space from the placenta rather than persistence of villous islands, which might occasionally appear globular with narrow zones of connection to the main portion of the placenta, like a succenturiate lobe. The twin peak sign cannot occur in monochorionic placentation because the single continuous chorion serves as an impenetrable barrier to the growth of the placental villi. These villi are thus excluded from extension into the potential interamniotic space of the monochorionic, diamniotic twin membrane (Fig. 68). The presence of placental villi in the interchorionic space of the dichorionic, diamniotic twin membrane can easily be identified by gross inspection of the placenta at delivery (Fig. 7), It is important to note that in the gross specimen (as in the illustrated case) and on sonograms during pregnancy, the interchorionic extension of placental villi usually appears focally scattered along the placental surface origin of the intertwin membrane, rather than occurring as a continuous ridge. In seeking the twin peak sign by ultrasonography, it is therefore imperative that the zone of intersection of the membrane with the placenta be carefully scruti~ nized over its entire length (or over as much as can be
A
A
1
B Figure 6 Chorionic and amniotic relationships explain the occurrence of the twin peak phenomenon only in dichorionic twin pregnancies. A, In a dichorionic pregnancy with fused placentas, both the amnions (A) and the chorions (C) reflect away from the placental surface at the point of origin of the septum. This creates a potential space in direct continuity with the chorionic villi and into which they can extend. B, In a monochorionic twin pregnancy, the septum is formed by reflection of the two amnions away from the placenta. There is a continuous single chorion, which provides an intact barrier, preventing extension of placental. villi into the potential interamniotic space.
detected, if portions are obscured by overlying fetal parts). It is important to recognize that absence of the twin peak sign does not imply that the pregnancy is mono· chorionic. When no chorionic peak is identified, the chorionid ty of the pregnancy remains indeterminate by this criterion. This is analogous to the decision pattern of identifying a twin pregnancy with two
J Ultrasound Med 11:571 - 577, 1992
Figure 7 Fused dichorionic twin placenta specimen after delivery. The segmental presence of placental villi are easily seen within the twin septum. This is the anatomic basis of the sonographic twin peak sign. (Photograph courtesy of Kurt Benirschke, MD, Professor of Pathology, University of California, San Diego.)
distinct placentas as dichorionic while the chorionicity of twins with a single placental zone cannot be distinguished, except by other criteria. As noted in the methodology, the data collected did not permit analysis for the frequency of this sign in dichorionic, fused-placenta pregnancies. Nonetheless, by comparison with observations of other signs, including membrane thickness and discordant sex twin pairs, it appeared that twin peaks are detectable in the majority of such cases when a diligent search for their presence is made. A possible pitfall in using chorionic peaks to assign chorionicity to triplet pregnancies should be noted. If three placentas have fused as adjacent bars or strips, identification of one peak between triplets A and B and a second peak between B and C is sufficient for confirming trichorionic pregnancy (Fig. SA). If, instead, the placentas are arranged radially, like pie wedges, peaks occurring between triplets A and B and between B and C would prove only that triplet B has a separate placenta from A and C, whereas triplets 'A and C could still share a monochorionic placenta (Fig. SB). The geometry of fused triplet placentas may be complex and not fully assessable sonographically. A second potential pitfall could occur if a uterine synechia was present, indenting the gestational sac of a monochorionic twin pregnancy in the region of the placenta. In over half of the cases in which a synechia is seen in pregnancy the placenta extends to, wraps around, or grows out along the adhesion. 1 This can create a peak-like elevation of the placenta which, in a twin pregnancy, could simulate the twin peak sign. It would, however, have no relationship to the intertwin membrane, except by chance proximity.
FINBERG
575
A
c
B
A B Figure 8 Variations in the geometry of fused triplet placentas. A, The three placental components are arranged as parallel bars or strips. Presence of twin peaks both between triplets A and B and between triplets B and C confirms the pregnancy to be trichorionic. 8, The three placental components are arranged radially. In this geometric arrangement, twin peaks between triplets A and B and between triplets B and C confirm only that B has a separate placenta from A and C. No statement about the placental interface between A and C can be made. These two may share a monochorionic placenta or have fused dichorionic placentas.
The twin peak phenomenon can occur at the margin of a placenta in a twin pregnancy with two distinct placentas if the intertwin membrane happens to originate at this site. (One such case, not included in this series, was seen.) Such an observation is not generally
576
THE TWIN PEAK SIGN
of diagnostic import. If there were potentia] concern that two separate placental zones represented a monochorionic twin placenta with a large succenturiate lobe, as in a case of significant size discordance of like· sex twins, presence of a twin peak would effectively exclude that possibility. The widespread interest in determining the type of placental development in twin pregnancies is predi+ cated on the substantially greater risk of shared-placenta twinning. Benirschke and Kim in 1973 2 reported a perinatal mortality of 26% in diamniotic, monochorionic twins as compared to 9% in diamniotic, dichorionic twins. (The mortality rate in the fortunately uncommon monoamniotic, monochorionic twin pregnancies was SO%.) Beginning with Barss and colleagues in 1985, recent efforts at distinguishing monochorionic from dichorionic twinning when separate placentas are not seen have focused on categorizing the twin septum as thin (monochorionic) or thick (dichorionic). 3 Hertzberg and coworkers found that accuracy in using septum thick· ness diminished with advancing pregnancy stage! Townsend and associates reported an 83% predictive value for both monochorionic and dichorionic twin pregnancies using membrane thickness criteria:" D' AJton and Dudley6 correctly determined chorionicity in 68 of 69 twin pregnancies by counting the number of layers in the intertwin membrane as seen on the son· agraphic image. Two layers indicated a monochorionic pregnancy, and three or four layers, a dichorionic set of twins. 6 In our own laboratory, we have not been able to verify this observation. The twin peak sign described in this report has received little if any attention in the ultrasound literature. A brief report by Bessis and Papiernik in 198t1 mentioned a '"lambda sign" as present in 20 of 24 dichorionic pregnancies and in none of six monochorionic twin pregnancies. They do not describe the find ing but have two labeled illustrations, one at 10 weeks and one at 32 weeks, that show a finding compatible with that described in the present report. Determination of the Type of Twinning To determine the type of twinning that has occurred, the following decision cascade is suggested. 1. At 10 weeks or Jess, count the number of gesta• tional sacs. Presence of two sacs, each with a live fetus, indicates dichorionic twinning. A single sac within which two live fetuses are present is monochorionic. Beyond 10 weeks the following order of criteria can be used. 2. Determine the gender of the twins. A male-female pair proves a dizygotic and, therefore, a dichorionic pregnancy. Uke sex or unidentified gender twins are of indeterminate chorionicity.
J Ultrasound Med 11:571- 577, 1992
3. Determine the number of separate placentas. Two distinct placentas indicate a dichorionic pregnancy. A single placental region is indeterminate, occurring in either a monochorionic or a fused dichorionic pregnancy. 4. With one placental zone seen, search for a twin peak sign. If it is present, the pregnancy has a dichorionic fused placenta. If none is detected, the chorionidly remains undetermined. 5. If the previous criteria have not identified the chorionic status, evaluate the intertwin membrane. A thick membrane indicates a high probability of dicho· rionic twinning, whereas a thin and wispy membrane implies a simi1arly high probability of a monochorionic twin pair. 6. If no membrane is detected, a monoamniotic, monochorionic twin pregnancy must be considered. The septum (usually of a diamniotic, monochorionic set of twins) may occasionally be present but undetected because of technical factors 8 or as a result of oligohydramnios within the sac of one twin so that the septum is so closely applied to this twin that it is not visualized . Such a twin is generally held tightly against the uterine wall, and is referred to as a •stuck twin ." This phenomenon can be demonstrated by having the mother roll so that this twin is nondependent and observing that the twin does not shift downward with gravity. Presence of entangled umbilical cords can con• firm monoamniotic twinning. 9 If these findings are inconclusive, single sac amniography with computed tomographic detection of intestinal contrast material in both twins can be used to prove that the pregnancy is monoamniotic, monochorionic. 111
REFERENCES 1. Finberg HJ: Uterine synechiae in pregnancy: Expanded criteria for recognition and review of clinical significance in 28 cases. J Ultrasound Med 10:547, 1991 2. Benirschke K, Kim CK: Multiple pregnancy (first of two parts). N Eng! J Med 288:1276, 1973 3. Barss VA, Benacerraf BR, Frigoletto FD: Ultrasonographic determination of chorion type in twin gestation. Obstet Gynecol66:779, 1985 4. Hertzberg BS, Kurtz AB, Choi HY, et al: Significance of membrane thickness in the sonographic evaluation of twin gestations. AJR 148:151, 1987 5. Townsend RR, Simpson GF, Filly RA: Membrane thickness in ultrasound prediction of chorionicity of twin gestations. J Ultrasound Med 7:327, 1988 6. D'Alton ME, Dudley DK; The ultrasonographic prediction of chorionicity in twin gestation. Am J Obstet Gy• necol160:557, 1989
J Ultrasound Med 11:571-577, 1992
7. Bessis R, Papiemik E: Echographic imagery of amniotic membranes in twin pregnancies. In Twin Research 3: Twin Biology and Multiple Pregnancy. New York, Alan R. Liss, 1981, p 183 8. Mahony 85, Filly RA, Callen PW: Amnionicity and chorionicity in twin pregnancies: Prediction using ultrasound. Radiology 155:205, 1985
FINBERG
577
9. Nyberg DA, Filly RA, Golbus MS, et al: Entangled urn• bilical cords: A sign of monoamniotic twins. j Ultrasound Med 3:29, 1984
10. Finberg HJ, Clewell WH: Definitive prenatal diagnosis of monoamniotic twins: Swallowed amniotic contrast agent detected in both twins on sonographically selected CT images. J Ultrasound Med 10:5123, 1991
