- Email: [email protected]
Ultrasound prediction of follicle volume: is the mean diameter reflective?*
FERTILITY AND STERILITY
Vol. 62, No.6, December 1994 Printed on acid~free paper in U. S. A.
Copyright 0 1994 The American Fertility Society
Ultrasound prediction of follicle volume: is the mean diameter reflective~*
Alan S. Penzias, M.D.t Adelina M. Emmi, M.D. Anil K. Dubey, Ph.D.
Lawrence C. Layman, M.D. Alan H. DeCherney, M.D. Richard H. Reindollar, M.D.
Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, Tufts University School of Medicine, Boston, Massachusetts
Objective: To evaluate the relationship between 2 dimensional sonographic measurement of ovarian follicles and their actual volume. Design: Prospective clinical study. Setting: The in vitro fertilization (IVF) program of a University based, tertiary care hospital. Patients and interventions: Sonographic categorization by shape, and measurement of 96 individual ovarian follicles immediately prior to aspiration for IVF. Each follicle was aspirated under direct ultrasound guidance and the volume recorded. The 96 follicles were visualized in a total of 14 patients from whom 2 to 27 oocytes were obtained. Main outcome measure: Total volume of each follicle. Results: Round and polygonal follicles exhibited a highly significant relationship between sonographically measured mean diameter and total follicle volume. The volume of follicles that were categorized as ellipsoid was not predicted by measurement of the longest diameter, shortest diameter or mean diameter. Conclusion: The mean diameter of round and polygonal follicles accurately predicts total follicular volume. However, clinical decisions in ovulation induction should be modified when the follicle shape is predominantly ellipsoid because the traditionally held belief that the sonographic measurement of the follicular diameter correlates with the follicular volume does not apply in those circumFertil Steril1994;62:1274-1276 stances. Key Words: Ovulation induction, ultrasound, fertilization in vitro
Follicle size has long been used as an indicator of oocyte maturity in ovulation induction and assisted reproductive technologies. The long-standing assumption has been that the two-dimensional measurement of a follicle by ultrasound (US) correlates with oocyte preparedness (1, 2). Presumably, therefore, the two-dimensional image of the follicle depicted by US corresponds well with the three-
Received January 26, 1994; revised and accepted July 1, 1994. * Presented at the Conjoint Meeting of The American Fertility Society and the Canadian Fertility and Andrology Society, Montreal, Quebec, Canada, October 11 to 14, 1993. t Reprint requests: Alan S. Penzias, M.D., New England Medical Center, 750 Washington Street, Box 36, Boston, Massachu· setts 02111 (FAX: 617-956-4202).
1274
Penzias et al. Communications-in-brief
dimensional structure and, hence, its total volume (3). We chose to evaluate the ability of sonographic follicle dimensions to predict total follicular volume. MATERIALS AND METHODS Patients
Ninety-six follicles were studied at the time of oocyte retrieval immediately before transvaginal aspiration. The vaginal transducer was positioned so that the roundest possible follicular image was obtained. The longest and shortest diameters of the follicle were measured and recorded. The follicles were classified by shape as either round, polygonal, Fertility and Sterility
l
or ellipsoid. The ovum capture needle was then guided into the follicle and its complete collapse observed. The contents of the each separate follicle were aspirated directly into a 15-mL conical tube and the volume recorded. Statistics
The expected volume of each follicle was calculated using the formula for the volume of a sphere, 11rr3 • For follicles that appeared ellipsoid, several values ofr, based on the longest, shortest, and mean diameters, were substituted in an attempt to find the best fit model. Because ovarian follicles may not be perfect spheres, a modified formula for the volume of nonspherical objects also was applied in an attempt to better predict the actual volume. Linear regression analysis was performed to explore the relationship between each of these calculated values and the true volume. In addition, a linear regression model was created to further explore the relationship of the actual measurements, longest, shortest, and mean diameters, with actual follicle volume. Therefore, a total of six predictor variables (3 calculated volumes using formulas for spheres and ellipsoids, 2 measured dimensions, and 1 calculated dimension) were compared with the actual volume obtained at retrieval. Statistical significance was assumed at P < 0.05.
Table 1 Relationship Between Measured Follicle Dimensions With Actual Follicle Volume r
Probability
0.548 0.551 0.670
<0.0005 <0.0005 <0.0005
0.635 0.739 0.766
<0.0005 <0.0005 <0.0005
0.320 0.013 0.252
0.118* 0.945* 0.224*
Variable All follicles (n = 96) Longest diameter Shortest diameter Mean diameter Round and polygonal follicles (n = 67) Longest diameter Shortest diameter Mean diameter Ellipsoid follicles (n = 29) Longest diameter Shortest diameter Mean diameter
* Not statistically significant.
relationship between the actual follicular volume obtained and each of six predictor variables. When the round and polygonal follicles were grouped together, the mean diameter was most linearly related to actual volume (r = 0.766; P < 0.0005). When ellipsoid follicles were examined as a group, none of the six variables was found to be predictive of actual volume. In fact, the variables overestimated follicle volume as often as they underestimated the follicle volume.
DISCUSSION RESULTS
Ninety-six follicles were studied in 14 individuals from whom 2 to 27 oocytes were obtained. A total of 67 follicles were categorized as round or polygonal and 29 as ellipsoid. For follicles that were categorized as round or polygonal, the mean values (±SEM) for long diameter, short diameter, and mean diameter were 18.55 ± 0.53, 16.33 ± 0.53, and 17.26 ± 0.48 mm, respectively (range, 14%). For ellipsoid follicles, the mean values (± SEM) for long diameter, short diameter, and mean diameter were 21.04 ± 0.62, 14.31 ± 0.76, and 17.16 ± 0.53 mm, respectively (range, 47%). Although the values for mean diameter of round and polygonal follicles and ellipsoid follicles were nearly identical, the internal ranges were markedly discrepant. Table 1 summarizes the relationship between follicle measurements and actual fluid volume for the whole population of follicles (n = 96) as well as its subsets after categorization by follicle shape. When all 96 follicles were analyzed together without regard to shape, we observed a significant linear Vol. 62, No.6, December 1994
In this study, we sought to identify predictors of ovarian follicular volume. We demonstrated that for round and polygonal follicles, the mean diameter determined by two-dimensional US accurately predicted the total volume. However, the volume of ellipsoid follicles was overestimated as often as it was underestimated by sonographic measurement. Because previous reports have described a relationship between the sonographic size of ovarian follicles and follicle volume, our findings have a direct bearing on this currently held belief. In a study by O'Herlihy et al. (4), 39 preovulatory follicles were examined in 36 patients (single follicles in 33 patients and 2 follicles in 3 others). When the sonographic mean diameter was compared with the calculated diameter derived from the actual volume, they noted a highly significant linear relationship (r = 0.847; P < 0.0001). In our study, we found that rounded follicles tended to be most prevalent in patients with the fewest follicles. It is likely that the follicles described by O'Herlihy (4) tended toward roundness. Mantzavinos et al. (5) evaluated 51 Penzias et al.
Communications-in-brief
1275
patients undergoing laparoscopic oocyte retrieval in hMG-stimulated IVF cycles and also reported a linear relationship between actual and calculated diameter (r = 0.7). However, this study is limited by the fact that the last US was performed 1 day before laparoscopy so that a direct correlation between specific follicles and sonographic measurements cannot be assured; also, the number of follicles examined per patient was small, again making it more likely that most follicles were round. Neither report commented specifically on the shape of the follicles studied nor stratified their analysis based on it. We elected to classify follicles as round, polygonal, or ellipsoid to make categorization as simple and practical as possible. We found that this uncomplicated system of classification was crucial in allowing its uniform application by different observers, especially in ovulation induction regimens incorporating GnRH agonist where there is a substantial increase in the number offollicles observed and oocytes retrieved compared with regimens employing hMG alone. Furthermore, we were able to demonstrate the lack of correlation between US measurements and actual volume for ellipsoid follicles as compared with round or polygonal follicles. Although the variable most reflective of the actual volume is the mean diameter, the volume of ellipsoid follicles cannot be predicted
1276
Penzias et al.
Communications-in-brief
with certainty by a two-dimensional image. It is possible that the measurement of a third dimension would improve the predictive value of ultrasonography. However, it would be difficult to assure that a third measurement is truly perpendicular to the index plane because there are no fixed reference points available within the follicle, and an accurate third measurement would be exceedingly difficult if not impractical in a crowded ovary. REFERENCES 1. Seibel MM, McArdle CR, Thompson IE, Berger MJ, Tay-
2.
3.
4.
5.
mor ML. The role of ultrasound in ovulation induction: a critical appraisal. Fertil Steril1981;36:573-7. Navot D, Margoliath EJ, Laufer N, Brzezinski A, Birkenfeld A, Schenker JG. Periovulatory 17 beta estradiol pattern in conception and nonconception cycles during menotropin treatment of anovulatory infertility. Fertil Steril 1987; 47:234-7. Simonetti S, Veeck LL, Jones HW Jr. Correlation offollicular fluid volume with oocyte morphology from follicles stimulated by human menopausal gonadotropin. Fertil Steril 1985;44:177-80. O'Herlihy C, De Crespigny LC, Lopata A, Johnston I, Hoult I, Robinson H. Preovulatory follicular size: a comparison of ultrasound and laparoscopic measurements. Fertil Steril 1980;34:24-6. Mantzavinos T, Garcia JE, Jones HW Jr. Ultrasound measurement of ovarian follicles stimulated by human gonadotropins for oocyte recovery and in vitro fertilization. Fertil Steril1983;40:461-5.
Fertility and Sterility
Vol. 62, No.6, December 1994 Printed on acid~free paper in U. S. A.
Copyright 0 1994 The American Fertility Society
Ultrasound prediction of follicle volume: is the mean diameter reflective~*
Alan S. Penzias, M.D.t Adelina M. Emmi, M.D. Anil K. Dubey, Ph.D.
Lawrence C. Layman, M.D. Alan H. DeCherney, M.D. Richard H. Reindollar, M.D.
Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, Tufts University School of Medicine, Boston, Massachusetts
Objective: To evaluate the relationship between 2 dimensional sonographic measurement of ovarian follicles and their actual volume. Design: Prospective clinical study. Setting: The in vitro fertilization (IVF) program of a University based, tertiary care hospital. Patients and interventions: Sonographic categorization by shape, and measurement of 96 individual ovarian follicles immediately prior to aspiration for IVF. Each follicle was aspirated under direct ultrasound guidance and the volume recorded. The 96 follicles were visualized in a total of 14 patients from whom 2 to 27 oocytes were obtained. Main outcome measure: Total volume of each follicle. Results: Round and polygonal follicles exhibited a highly significant relationship between sonographically measured mean diameter and total follicle volume. The volume of follicles that were categorized as ellipsoid was not predicted by measurement of the longest diameter, shortest diameter or mean diameter. Conclusion: The mean diameter of round and polygonal follicles accurately predicts total follicular volume. However, clinical decisions in ovulation induction should be modified when the follicle shape is predominantly ellipsoid because the traditionally held belief that the sonographic measurement of the follicular diameter correlates with the follicular volume does not apply in those circumFertil Steril1994;62:1274-1276 stances. Key Words: Ovulation induction, ultrasound, fertilization in vitro
Follicle size has long been used as an indicator of oocyte maturity in ovulation induction and assisted reproductive technologies. The long-standing assumption has been that the two-dimensional measurement of a follicle by ultrasound (US) correlates with oocyte preparedness (1, 2). Presumably, therefore, the two-dimensional image of the follicle depicted by US corresponds well with the three-
Received January 26, 1994; revised and accepted July 1, 1994. * Presented at the Conjoint Meeting of The American Fertility Society and the Canadian Fertility and Andrology Society, Montreal, Quebec, Canada, October 11 to 14, 1993. t Reprint requests: Alan S. Penzias, M.D., New England Medical Center, 750 Washington Street, Box 36, Boston, Massachu· setts 02111 (FAX: 617-956-4202).
1274
Penzias et al. Communications-in-brief
dimensional structure and, hence, its total volume (3). We chose to evaluate the ability of sonographic follicle dimensions to predict total follicular volume. MATERIALS AND METHODS Patients
Ninety-six follicles were studied at the time of oocyte retrieval immediately before transvaginal aspiration. The vaginal transducer was positioned so that the roundest possible follicular image was obtained. The longest and shortest diameters of the follicle were measured and recorded. The follicles were classified by shape as either round, polygonal, Fertility and Sterility
l
or ellipsoid. The ovum capture needle was then guided into the follicle and its complete collapse observed. The contents of the each separate follicle were aspirated directly into a 15-mL conical tube and the volume recorded. Statistics
The expected volume of each follicle was calculated using the formula for the volume of a sphere, 11rr3 • For follicles that appeared ellipsoid, several values ofr, based on the longest, shortest, and mean diameters, were substituted in an attempt to find the best fit model. Because ovarian follicles may not be perfect spheres, a modified formula for the volume of nonspherical objects also was applied in an attempt to better predict the actual volume. Linear regression analysis was performed to explore the relationship between each of these calculated values and the true volume. In addition, a linear regression model was created to further explore the relationship of the actual measurements, longest, shortest, and mean diameters, with actual follicle volume. Therefore, a total of six predictor variables (3 calculated volumes using formulas for spheres and ellipsoids, 2 measured dimensions, and 1 calculated dimension) were compared with the actual volume obtained at retrieval. Statistical significance was assumed at P < 0.05.
Table 1 Relationship Between Measured Follicle Dimensions With Actual Follicle Volume r
Probability
0.548 0.551 0.670
<0.0005 <0.0005 <0.0005
0.635 0.739 0.766
<0.0005 <0.0005 <0.0005
0.320 0.013 0.252
0.118* 0.945* 0.224*
Variable All follicles (n = 96) Longest diameter Shortest diameter Mean diameter Round and polygonal follicles (n = 67) Longest diameter Shortest diameter Mean diameter Ellipsoid follicles (n = 29) Longest diameter Shortest diameter Mean diameter
* Not statistically significant.
relationship between the actual follicular volume obtained and each of six predictor variables. When the round and polygonal follicles were grouped together, the mean diameter was most linearly related to actual volume (r = 0.766; P < 0.0005). When ellipsoid follicles were examined as a group, none of the six variables was found to be predictive of actual volume. In fact, the variables overestimated follicle volume as often as they underestimated the follicle volume.
DISCUSSION RESULTS
Ninety-six follicles were studied in 14 individuals from whom 2 to 27 oocytes were obtained. A total of 67 follicles were categorized as round or polygonal and 29 as ellipsoid. For follicles that were categorized as round or polygonal, the mean values (±SEM) for long diameter, short diameter, and mean diameter were 18.55 ± 0.53, 16.33 ± 0.53, and 17.26 ± 0.48 mm, respectively (range, 14%). For ellipsoid follicles, the mean values (± SEM) for long diameter, short diameter, and mean diameter were 21.04 ± 0.62, 14.31 ± 0.76, and 17.16 ± 0.53 mm, respectively (range, 47%). Although the values for mean diameter of round and polygonal follicles and ellipsoid follicles were nearly identical, the internal ranges were markedly discrepant. Table 1 summarizes the relationship between follicle measurements and actual fluid volume for the whole population of follicles (n = 96) as well as its subsets after categorization by follicle shape. When all 96 follicles were analyzed together without regard to shape, we observed a significant linear Vol. 62, No.6, December 1994
In this study, we sought to identify predictors of ovarian follicular volume. We demonstrated that for round and polygonal follicles, the mean diameter determined by two-dimensional US accurately predicted the total volume. However, the volume of ellipsoid follicles was overestimated as often as it was underestimated by sonographic measurement. Because previous reports have described a relationship between the sonographic size of ovarian follicles and follicle volume, our findings have a direct bearing on this currently held belief. In a study by O'Herlihy et al. (4), 39 preovulatory follicles were examined in 36 patients (single follicles in 33 patients and 2 follicles in 3 others). When the sonographic mean diameter was compared with the calculated diameter derived from the actual volume, they noted a highly significant linear relationship (r = 0.847; P < 0.0001). In our study, we found that rounded follicles tended to be most prevalent in patients with the fewest follicles. It is likely that the follicles described by O'Herlihy (4) tended toward roundness. Mantzavinos et al. (5) evaluated 51 Penzias et al.
Communications-in-brief
1275
patients undergoing laparoscopic oocyte retrieval in hMG-stimulated IVF cycles and also reported a linear relationship between actual and calculated diameter (r = 0.7). However, this study is limited by the fact that the last US was performed 1 day before laparoscopy so that a direct correlation between specific follicles and sonographic measurements cannot be assured; also, the number of follicles examined per patient was small, again making it more likely that most follicles were round. Neither report commented specifically on the shape of the follicles studied nor stratified their analysis based on it. We elected to classify follicles as round, polygonal, or ellipsoid to make categorization as simple and practical as possible. We found that this uncomplicated system of classification was crucial in allowing its uniform application by different observers, especially in ovulation induction regimens incorporating GnRH agonist where there is a substantial increase in the number offollicles observed and oocytes retrieved compared with regimens employing hMG alone. Furthermore, we were able to demonstrate the lack of correlation between US measurements and actual volume for ellipsoid follicles as compared with round or polygonal follicles. Although the variable most reflective of the actual volume is the mean diameter, the volume of ellipsoid follicles cannot be predicted
1276
Penzias et al.
Communications-in-brief
with certainty by a two-dimensional image. It is possible that the measurement of a third dimension would improve the predictive value of ultrasonography. However, it would be difficult to assure that a third measurement is truly perpendicular to the index plane because there are no fixed reference points available within the follicle, and an accurate third measurement would be exceedingly difficult if not impractical in a crowded ovary. REFERENCES 1. Seibel MM, McArdle CR, Thompson IE, Berger MJ, Tay-
2.
3.
4.
5.
mor ML. The role of ultrasound in ovulation induction: a critical appraisal. Fertil Steril1981;36:573-7. Navot D, Margoliath EJ, Laufer N, Brzezinski A, Birkenfeld A, Schenker JG. Periovulatory 17 beta estradiol pattern in conception and nonconception cycles during menotropin treatment of anovulatory infertility. Fertil Steril 1987; 47:234-7. Simonetti S, Veeck LL, Jones HW Jr. Correlation offollicular fluid volume with oocyte morphology from follicles stimulated by human menopausal gonadotropin. Fertil Steril 1985;44:177-80. O'Herlihy C, De Crespigny LC, Lopata A, Johnston I, Hoult I, Robinson H. Preovulatory follicular size: a comparison of ultrasound and laparoscopic measurements. Fertil Steril 1980;34:24-6. Mantzavinos T, Garcia JE, Jones HW Jr. Ultrasound measurement of ovarian follicles stimulated by human gonadotropins for oocyte recovery and in vitro fertilization. Fertil Steril1983;40:461-5.
Fertility and Sterility