Chapter 10: Persistence / Late Disappearance

---

The criterion of late disappearance or persistence of choroid plexus cysts beyond a nominated gestational age was developed by Chitkara⁴⁷ from the early data of Nicolaides¹³¹, Ricketts¹⁵⁹ and the report by Chudleigh⁴⁹ that in her series of five normal fetuses all cysts had resolved by 23 weeks. 

10.1 Early “evidence” of persistence

Two fetuses in Nicolaides’ 1986 series of trisomic fetuses were at 24 weeks gestation and one was terminated prior to this age (preventing the assessment of persistence). Rickett’s only case (T21) was discovered at 23 weeks. Therefore in the data up to that date three of four trisomic fetuses with CPC were at least 23 wks gestation. Ricketts pointed this out but did not draw any conclusions. Chudleigh’s report of five normal fetuses in which the cysts resolved by 23 weeks was the only population with which to make comparisons. The association with a delay in the disappearance of CPC was inevitable. In fact no relationship between disappearance and normality, or non-disappearance and abnormality had been established at this point.

Chitkara’s paper in 1988 was the first relatively large analytical series, with 41 cases of CPC. Only one fetus in this series was trisomic (T18), a twin. Amongst her 40 normal cases six had persistent cysts beyond 24 weeks, three of these persisting beyond 28 weeks. A seventh case of CPC was only diagnosed at 33 weeks. In her only trisomic fetus the cyst was present at 22 weeks prior to selective feticide. In spite of the fact that 7/40 cysts were seen after 24 weeks in normal fetuses and that the trisomic fetus was not scanned after 22 weeks, Chitkara advised her readers to be wary of persistent cysts. Chitkara reviewed the data from previous investigators and found four out of eight cases of trisomy to support the conclusion that:

“Chromosomal studies should also be considered and are strongly recommended whenever associated anomalies are detected or when the cysts are large, bilateral and persistent after 20-22 weeks’ gestation.” (Italics mine.)⁴⁷

The only conclusion that could possibly be drawn from her own data (as shown in Table 14) is that persistence of cysts is common in non-trisomic fetuses. Chitkara seems to have decided to look elsewhere for her conclusions, and placed the emphasis on the opposite finding, drawn from her literature review rather than her own data. Already the paradigm is exerting it's effects on investigators' thinking. Unfortunately, her conclusion is quoted or paraphrased extensively in many of the articles which were subsequently produced, and became part of the paradigm in the clinical world.

Chitkara et al	Number	Present at or after 24 wks
Normal	40	7
Trisomy 18	1	selective feticide at 22 wks

Table 14: Cases of persistence, from data in Chitkara⁴⁷.

There were authors trying to establish a more reasoned and scientific basis for opinion. An autopsy study of three cases (one with T18) by Farhood⁷⁵ in 1987 pointed out the relationship of the cysts to the histogenesis of the choroid, as did Fitzsimmons⁷⁹. In 1988, DeRoo⁶² also referred to the anatomical work of Shuangshoti described in Chapter 3.

10.2 The Need for Reassurance

Camurri⁴⁰ interpreted Chudleigh as implying “that cysts were a normal variant because of their disappearance between 20 and 23 weeks”⁴⁰ (Emphasis mine).

It had seemed to some that the presence of a cyst suggested the possibility of an intrinsic abnormality in the brain. Disappearance therefore meant that the brain was developing normally. Cases of obstructive hydrocephalus from large cysts situated in the third ventricle have been reported in adults, and hydrocephalus has been noted in several fetuses. These “colloid” cysts in adults probably have a different pathogenesis from those commonly seen in fetuses.

In early 1989 Hershey¹⁰⁰ felt unable to justify delaying action until a repeat scan was performed. He suggested that amniocentesis should be offered at the time of cyst discovery rather than as Chitkara suggested, waiting until after 20-22 weeks to see whether they disappear. In response to this suggestion, Chitkara advocated the second scan:

“...as it is reassuring to know that the chance of a chromosomal abnormality is small if the cysts are small at the time of the initial examination and have decreased further in size by 20-22 weeks”

Chitkara¹⁰⁰. (In reply to letter from Hershey.)

Unfortunately neither Chitkara’s nor anyone else’s data actually supports such reassurance. Other authors (Ostlere¹⁴⁶, Platt¹⁵²) also suggested a re-scan at 20-22 weeks. The purpose of this scan is not apparently to document disappearance although many would be gone by this time, but to note any decrease in size. Benacerraf¹² suggested a repeat scan would be better utilzed as another chance for the evaluation of fetal anatomy. Oettinger¹⁴² and Nava¹²⁵ suggested a repeated scan in cases where amnio was refused, but not specifically as an anatomy re-check, rather to check for IUGR and other late signs of T18.

Hopes for reassurance was not given support in 1992 when Rotmensch¹⁶³ reported a case of T21 in a fetus with bilateral CPC as the only ultrasound detected abnormality despite an apparently highly detailed assessment. The cysts had regressed in size from 10mm at 16 weeks to 5mm at 19 weeks, immediately prior to pregnancy termination. Rotmensch concludes that “diminishing cyst size alone should not be considered sufficient reassurance about the normality of the karyotype”.

10.3 How Late is Late?

There is difficulty establishing the fetal age which constitutes “late” disappearance. In Perpignano’s¹⁵⁰ 1992 series of 87 cases, the mean age of disappearance of cysts was 25.4 weeks, with the 2 aneuploid fetuses conforming to that average. Do these gestations represent the actual time at which the cysts spontaneously resolved or are they the gestation at which a scan was done which no longer showed cysts? Normal fetuses had cysts persisting up to 36 weeks, in agreement with the data of Chitkara⁴⁷. Thus the initial suggestion by Chitkara that 24 weeks (she later suggested 22wks) should be the cut-off is not supported by the natural history of cysts in the normal population, and by uncertainty of the actual gestation at resolution.

10.4 Lack of Concurrence in the Literature

DeRoo⁶² in 1988 was the first to propose “that serial sonography for cyst evaluation is not useful in determining fetal prognosis”, thus challenging the school that was forming. In a thorough literature review in 1993, Kennedy¹¹⁰ pointed out the major weaknesses of the persistence theory: many series have shown cyst persistence in normal fetuses and they are a frequent incidental finding in autopsies. When in 1989 Khouzam¹¹¹ reported a case study in which a T18 fetus had a cyst persisting beyond 36 weeks, Lodeiro, as if to counter this, reported a case in which a cyst in a normal fetus persisted until 34 weeks.

In reviewing the literature concerning cyst persistence Lodeiro pointed out that among the previous investigators their “findings did not seem to concur”. This lack of concurrence seems to be due to authors reiterating the suggestions of some of the early investigators, strengthening the paradigm, rather than thinking about their own data. Zerres²⁰⁰, for example, in a 1992 report on 25 cases of CPC provided no information about cyst size, the fetal age when the cysts were discovered or if there was persistence in his own study. Despite this lack of supporting evidence from within the article, Zerres comes to the conclusion that these “cysts are often larger and persist beyond the 24^th week of gestation”. The readers of this article cannot confirm if its conclusions are justified or not, as the data are lacking. 

The largest series to date, from Nadel¹²³ (234 cysts), found three normal fetuses with cysts persisting into the third trimester. Three other cysts in the series were discovered initially in fetuses 24 weeks or older, two having trisomy18, one being normal. Although this constitutes a 33% association, no comment on the relation between persistence or late detection and aneuploidy was made in the study.

10.5 Statistical Analysis

From the literature survey, 997 fetuses could be identified where the gestational age at cyst disappearance was either stated explicitly or could be surmised from the data provided. 110 of these were trisomic or had a sex chromosome aberration. At least 45 of the normal fetuses (sensitivity 94.7%) were described with cyst persistence beyond 23 weeks, but many of the cysts in normal fetuses in the literature were not followed serially. Similarly in 15 aneuploid fetuses (15.8%) the cysts were present at 23 weeks or after, but those fetuses that were terminated for T18 earlier than 23 weeks obviously could not be evaluated. 

Those authors who provide enough information about cyst persistence to allow the statistical analysis are shown in Table 15, below.

A total of 1005 cases are extant, of which only 13 aneuploid fetuses had CPC that persisted beyond 23 weeks. In contrast 43 CPC in normal fetuses persisted. Fig 14 shows the graph of persistence in both euploid and aneuploid fetuses, and demonstrates quite clearly that this is a relatively rare occurrence, giving less credence to its potential use as a discriminating factor. When these figures are displayed in a 2x2 table (Table 16), the sensitivity is noted to be particularly low, while the specificity is reasonably good.

Author	No. Cysts	Euploid > 23wks	Aneuploid (T18)	Aneuploid >23 (T18)
Chudleigh^49	5	1
Nicolaides^131	4		2	2
Ricketts^159	4	1	1	1
Ostlere^144	11	1
Chitkara^47	41	7	1
Clark^50	5	2
Hertzberg^102	31	2
Benacerraf^14	38	10
Fitzsimmons^79	5		5
Gabrielli^80	82	1	4
Chan^45	13	1
Khouzam^111	1		1	1
Lodeiro^120	1	1
Benacerraf^12	5		5	1
Ostlere^146	100	4	3	2
Thorpe-Beeston^185	83	1	20
Chinn^46	38
Twining^189	19		2
Platt^152	71	8	4
Achiron^2	30	1	5
Rotmensch^163	1		1
Perpignano^150	87		6	2
Nadel^123	234	1	12	1
Porto^154	63	1	6	1
Nyberg^136	11		11	1
Kennedy^110	22		3	1
Total	1005	43	94 (77)	13 (11)

Table 15: Articles which discuss persistence, and number of persisting > 23 wks.

Fig14: Persistence > 23 wks in T18 and Normal fetuses

	T18	Normal	PPV
Persistent >23 wks	11	43	23.2%
Not seen > 23 wks	66	869	Likelihood Ratio
Total	77	912	0.0304
Sensitivity	14.3%	95.3%	Specificity
False Negative Rate	85.7%	4.7%	False Positive Rate

Table 16: Persistence: 2 x 2 table.

Sensitivity, sensitivity, false positive rate, false negative rate, positive predictive value, and the likelihood ratio are calculated in Table 16.  Using the chi-squared test, the difference in persistence between aneuploid and normal fetuses shows statistically significance (p = 0.0021).  However, the sensitivity of 14.3% is particularly low, while the specificity is reasonably high, so again the question of clinical usefulness must be evaluated. Therefore, using Bayes Theorem to include the prevalence of T18 in the calculation, we find the equation:

Odds (Risk) of Trisomy 18 in a persistent cyst = 

0.0433% x 3.04% = 0.132% or 1 in 759.

The odds are much higher than the background risk for amniocentesis.  If this criterion were used to triage for amniocentesis nearly 4 normal fetuses would be lost for every one with T18 diagnosed.

For all its early trumpeting, the low risk means that persistence is useless in any individual case. It is disturbing to note how this criterion is still supported in recent articles and letters (Carmody²³⁸, Hershey¹⁰¹, Zerres²⁰⁰) reinforcing it as a part of the paradigm in the sonographic community and perhaps causing unnecessary delays in other more appropriate diagnostic or therapeutic procedures.

10.5 Conclusion

Sturla Eik-nes† , at the Annual Scientific Meeting of the Australian Society for Ultrasound in Medicine in Sydney, 1995, stated that any risk of aneuploidy is related to the CPC being present. If the cyst goes away, that does not erase the fact that it was there before, and the association with T18 remains. 

In summary, cyst persistence does not discriminate fetuses at higher risk for T18 for the following reasons:

· the natural history of CPC is not fully understood.

· as most cysts regress, there is a false sense of reassurance in aneuploid fetuses

· there is a high rate of persistence in normal fetuses.

· there is a requirement to extend possibly aneuploid pregnancies beyond the suitable and/or legal time for termination to see if this criterion is present.

· while there is a statistical significance to this criterion, the sensitivity is particularly low.

· adjustment of risk by Bayes Theorem shows the risk with persistent cysts is not sufficient to indicate amniocentesis in the median maternal ages group. 

This criterion does not assist in the discrimination of the fetus with T18. The modification of risk is elaborated for various maternal ages in the Appendix.

---