Chapter 8: Size


 

There has been no definitive longitudinal study of the ultrasound appearance of the natural history of CPC in normal and abnormal fetuses.  The history of CPC on ultrasound therefore relies in the anatomical studies of Shuangshoti172,173,174 and the many cross-sectional case reports at various ages of gestation.  Although many of these patients have had two or three scans, we do not know with certainty what happens to the size of individual cysts over an interval of a week or less.  The fact that the cysts are not often seen prior to fifteen weeks and are similarly uncommon after 24 weeks indicates that they do follow a cycle of growth and resolution through this period.  This would indicate, logically, that the cyst’s size will vary —  it will grow larger and eventually shrink again. The fact that the Geelong Hospital survey found no correlation with gestational age does not counteract this argument, as a linear correlation was being sought.  Bronshtein30 has described the case of a cyst developing from the “spongy” appearance of tubules at 14 weeks (seen with TV scanning) to a single cyst at 17 weeks.  This corresponds to the variation in tubule and cyst size and number in the stages of choroid plexus development described by Shuangshoti, discussed earlier.

8.1           Large size in early reports

The first article to mention the size of CPC was that of Bundy33 in 1986, in a review of 12 cases of T18.  One case with CPC was noted in the “Table of Sonographic Findings” but it was not discussed in the text of the article.  A scan showing the CPC was provided and the caption described them as  “Large choroid plexus cysts.  This is most likely a normal variant since they are seen in many other normal fetuses at around 20 weeks’ gestation.”  Any inferences about the significance of the size of CPC were drawn by readers of this article as the association was not explicitly made by Bundy.

 

The first article which directly implicated cyst size in the risk for T18 was that of Furness81 in 1987, in which multiple cysts “replaced the choroid bilaterally”.   One of the images in her article demonstrated several cysts evidently 10 mm or more in size (estimated from the centimetre markings on the image).  Two other fetuses in her series of 30 had T18 but the cysts in these fetuses were evidently small or less extensive.  In fact, the emphasis in this article was placed on the extensive nature of the cysts in one case rather than absolute size, and the fact that two out the three cases had small rather than large cysts was apparently editorially judged to be irrelevant.  This is evidenced by the fact that it was not even stated in the article that these were in T18 fetuses.

 

In 1988, Chitkara’s47 presented one case of T18 in 41 cases of CPC, in a twin with bilateral cysts of 20mm.  By this time in the literature, three fetuses out of eight with trisomy 18 were noted to have large cysts.  Two had smaller cysts, and in three cases the size was not mentioned.  This prompted Chitkara to draw the conclusion that large cysts were significant.  

 

Clark50 presented five cases of cysts in normal fetuses.  In three of these the size was greater than 10mm.

8.2           Changing size

Alteration in size has been detected in many cases when serial scans have been performed.  Diminution has been considered by some to be a good sign, providing reassurance of impending resolution, whereas increase in size has been considered a bad sign.  However, this merely reflects the natural history of cysts as discussed, and many authors have dismissed such suggestions as illogical.  Twining189 and Rotmensch163 have described cases in which cyst size either decreased or the cysts resolved in aneuploid fetuses.

8.3           Reviews of size

After her review of the literature, DeRoo62 felt that no conclusion should be drawn in regard to size as “conclusive data are lacking.”  As the number of reports increased, substantial variations in cyst size in both normal and aneuploid fetuses were being documented. 

 

DeRoo’s and Chan’s45 series had average size of cysts as 6.7mm and 6mm respectively.  Chinn felt that the average size of 4.5mm in his series indicated the importance of high resolution equipment which “is related to better detection of smaller cysts.”  In older articles and in those where equipment was not top quality, smaller cysts were undoubtedly missed.  Also due to the already growing paradigm, many smaller cysts of 2 –3 mm  were most likely ignored, thus increasing the average size in many articles. 

 

The purported association of size with aneuploidy became even less clear.  Nadel123 and Perpignano150 have independently reported on aneuploid fetuses with cysts as small as 3mm.  The largest reported cyst in a normal fetus was 21mm, and in an aneuploid fetus it was 30mm (associated with gross hydrocephalus).  The authors who have provided suitable information to analyse for cyst size are tabulated below.

 

 

 

Author

Euploid 

</= 10mm

Euploid >10mm

Aneuploid </= 10

Aneuploid > 10

Chudleigh49

5

 

 

 

Bundy33

 

 

 

1

Ostlere144

11

 

 

 

Furness81

 

 

 

1

Chitkara47

37

3

 

1

Clark50

2

3

 

 

DeRoo62

16

 

 

 

Gabrielli83

78

 

2

2

Chan45

13

 

 

 

Khouzam111

 

 

 

1

Lodeiro120

1

 

 

 

Camurri40

 

 

 

1

Ostlere145

93

4

1

2

Twining189

 

 

 

1

Achiron2

 

 

 

5

Rotmensch163

 

 

1

 

Perpignano150

 

 

4

2

Nadel123

 

 

11

2

Porto154

57

 

5

1

Kennedy110

 

 

3

 

Burrows34

 

 

1

 

Oettinger142

 

 

1

1

Gross91

65

12

3

 

Total

378

23

32 (29-T18)

21(20-T18)

Table 8: Authors who discuss cyst size and numbers of cysts above or below 10mm

 

8.4           Statistical Analysis

Cyst size is mentioned in 37 articles, but in some only the average sizes  of the cysts are given.  In the 23 articles where the absolute cyst size was given or could be surmised from the data, a total of 454 cases are available for evaluation of which 53 had chromosomal or autosomal abnormality.  Of the T18 fetuses, 29 were 10mm or less in their largest dimension, and 20 were 11mm or more.  In the normal fetuses 23 cases were 11mm or greater.

 

These figures are displayed in the graph (Fig12) and the 2x2 table below.

Fig 12: Cyst size above or below 1.0cm

 

 

 

T18

Normal

PPV

> 1.0 cm

20

23

46.5%

</= 1.0 cm

29

378

Likelihood ratio

Total

49

401

0.071 (7.1%)

Sensitivity

40.8%

94.3%

Specificity

False Negative Rate

59.2%

5.7%

False Positive Rate

Table 9: Cyst size: 2 x 2 table..

 

 

From this table the sensitivity of cyst size appears poor, while the specificity is good, due to the rarity of T18.  We can derive a positive predictive value (PPV) of 20/43 or 46.5% for cysts above 1.0cm in size, as nearly half of the cysts reported over 1.0 cm were in fetuses with T18.  The difference between the sensitivity and the specificity can be used to express the difference in odds of a large cysts being in either an euploid or T18 fetus.  This difference in odds is 40.8% - 94.3% or 0.534 (0.395 – 0.674, 5th% and 95th% confidence intervals.) The odds ratio is calculated to be 11.3 (that is 11.3 to 1, in betting terms) that a large cyst is more likely to be in a T18 fetus.  This of course assumes that reporting has been complete and no publishing bias is involved.  Using the chi-squared test, p was calculated to be less than 0.001, indicating a statistical significance in these methods of assessing the difference in cyst size between euploid and aneuploid fetuses.

 

In a clinical context however, decisions should not be based on the sensitivity of a test alone, nor even on the statistical significance of the results.  These figures fail to take into consideration the prevalence of T18 in the pregnant population.  Decisions should be based on the change in risk that occurs after a test is found to be positive.  Using Bayes Theorem to introduce the prevalence of T18 into the calculation, we find that the test is not as good as we might have thought at first.  The prevalence of T18 is multiplied by the likelihood ratio to produce the post-test risk. 

 

 

Odds (Risk) of T18 in a cyst >1.0 cm = 

 

0.0433% x 7.1% = 0.31%  or 1 in 325. 

 

This risk is lower than the 1/200 risk of fetal loss associated with amniocentesis.  If this criteria was used to triage for amniocentesis, the fetal loss would be 1.75 times the detection of T18, despite the earlier statistically derived odds, PPV and statistical significance.  From this method of calculation it can be seen than while larger cysts are  significantly associated with aneuploidy, they do not alter the risks enough to indicate amniocentesis in the median maternal age group.  Further breakdown by maternal age is provided in the Appendix.

8.6           Conclusion

In summary, the findings with the criterion of size are:

 

·        cyst size changes in normal and aneuploid fetuses according the cycle of plexus growth described by Shuangshoti. 

·        analysis shows that larger cysts are more common in aneuploid fetuses and that this difference is statistically significant.

·        however, adjustment of risk by Bayes Theorem shows that the risk with large cysts is not sufficient to indicate amniocentesis in the lower maternal ages. 

 

This criterion does not does not discriminate the patient at increased risk for aneuploidy in any individual case.