Predictive performance of BI-RADS magnetic resonance imaging descriptors in the context of suspicious (category 4) findings

Almeida, João Ricardo Maltez de; Gomes, André Boechat; Barros, Thomas Pitangueiras; Fahel, Paulo Eduardo; Rocha, Mário de Seixas

doi:10.1590/0100-3984.2015.0021

Acessibilidade / Reportar erro

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

ORIGINAL ARTICLES • Radiol Bras 49 (3) • May-Jun 2016 • https://doi.org/10.1590/0100-3984.2015.0021 copy

Predictive performance of BI-RADS magnetic resonance imaging descriptors in the context of suspicious (category 4) findings^* * Study conducted at the Clínica de Assistência à Mulher (CAM), Salvador, BA, Brazil.

Desempenho preditivo dos descritores de ressonância magnética do BI-RADS no contexto de achados suspeitos (categoria 4)

Authorship SCIMAGO INSTITUTIONS RANKINGS

Abstract

Objective:

To determine the positive predictive value (PPV) and likelihood ratio for magnetic resonance imaging (MRI) characteristics of category 4 lesions, as described in the Breast Imaging Reporting and Data System (BI-RADS^®) lexicon, as well as to test the predictive performance of the descriptors using multivariate analysis and the area under the curve derived from a receiver operating characteristic (ROC) curve.

Materials and Methods:

This was a double-blind review study of 121 suspicious findings from 98 women examined between 2009 and 2013. The terminology was based on the 2013 edition of the BI-RADS.

Results:

Of the 121 suspicious findings, 53 (43.8%) were proven to be malignant lesions, with no significant difference between mass and non-mass enhancement (p = 0.846). The PPVs were highest for masses with a spiculated margin (71%) and round shape (63%), whereas segmental distribution achieved a high PPV (80%) for non-mass enhancement. Kinetic analyses performed poorly, except for type 3 curves applied to masses (PPV of 73%). Logistic regression models were significant for both patterns, although the results were better for masses, particularly when kinetic assessments were included (p = 0.015; pseudo R² = 0.48; area under the curve = 90%).

Conclusion:

Some BI-RADS MRI descriptors have high PPV and good predictive performance-as demonstrated by ROC curve and multivariate analysis-when applied to BI-RADS category 4 findings. This may allow future stratification of this category.

Keywords:
Magnetic resonance imaging; Breast neoplasms; Predictive value of tests; Likelihood functions

	Lesions^* * Data in parentheses are percentages of the total number of masses (n = 56), except on the first row, where they are percentages of the total number of lesions (n = 121);	Positive lesions
Descriptor	N (%)	N	PPV^† † Expressed as %; —, Values that could not be calculated due to a lack of the finding in the study sample. [95% CI]	PLR^† † Expressed as %; —, Values that could not be calculated due to a lack of the finding in the study sample. [95% CI]
Mass	56 (46.28)	24	43 [30–57]	0.96 [0.65–1.42]
Shape
Oval	19 (33.93)	4	21 [6–46]	0.36 [0.14–0.94]
Round	16 (28.57)	10	63 [35–85]	2.22 [0.94–5.27]
Irregular	21 (37.50)	10	48 [26–70]	1.21 [0.62–2.38]
Margin
Circumscribed	7 (12.50)	1	14 [0.4–58]	0.22 [0.03–1.73]
Irregular	42 (75.00)	18	43 [28–59]	1.00 [0.74–1.36]
Spiculated	7 (12.50)	5	71 [29–96]	3.33 [0.71–15.7]
Internal enhancement
Homogeneous	26 (46.43)	7	27 [12–48]	0.49 [0.25–0.98]
Heterogeneous	17 (30.36)	10	59 [33–82]	1.90 [0.85–4.27]
Rim enhancement	13 (23.21)	7	54 [5–81]	1.56 [0.60–4.04]
Dark internal septations	0	0	—	—
T2 signal
Low	18 (32.14)	10	56 [31–79]	1.67 [0.78–3.58]
High	38 (67.86)	14	37 [22–54]	0.78 [0.53–1.15]

	Lesions^* * Data in parentheses are percentages of the total number of non-mass enhancement (n = 65), except on the first row, where they are percentages of the total number of lesions (n = 121).	Positive lesions
Descriptor	N (%)	N	PPV^† † Expressed as %. —, Values that could not be calculated due to a lack of the finding in the study sample. [95% CI]	PLR^† † Expressed as %. —, Values that could not be calculated due to a lack of the finding in the study sample. [95% CI]
Non-mass enhancement	65 (53.72)	29	45 [32–58]	1.03 [0.74–1.44]
Distribution and symmetry
Focal	25 (38.46)	7	28 [12–49]	0.48 [0.23–1.00]
Linear	11 (16.92)	4	36 [11–69]	0.71 [0.23–2.19]
Segmental	15 (23.08)	12	80 [52–96]	4.97 [1.55–15.90]
Regional	4 (6.15)	1	25 [63–81]	0.41 [0.05–3.77]
Multiple regions	1 (1.54)	1	100 [0.03–100]	—
Diffuse	0	0	—	—
Symmetric	0	0	—	—
Asymmetric	9 (13.85)	4	44 [14–79]	0.99 [0.29–3.37]
Internal enhancement
Homogeneous	8 (12.31)	1	13 [32–53]	0.18 [0.02–1.36]
Heterogeneous	54 (83.08)	25	46 [33–60]	1.07 [0.86–1.33]
Clumped	3 (4.62)	3	100 [29–100]	—
T2 signal
Low	47 (72.31)	21	45 [30–60]	1.00 [0.74–1.36]
High	18 (27.69)	8	44 [22–69]	0.99 [0.45–2.19]

	Lesions^* * Data in parentheses are percentages of the total number of lesions by main enhancement pattern: mass and non-mass enhancement.	Positive lesions
Type of curve	N (%)	N	PPV^† † Expressed as %. [95% CI]	PLR^† † Expressed as %. [95% CI]
Mass (n = 51)
Type 1	8 (15.69)	3	38 [0.09–76]	0.73 [0.20–2.74]
Type 2	28 (54.90)	9	32 [16–52]	0.58 [0.33–1.02]
Type 3	15 (29.41)	11	73 [45–92]	3.35 [1.23–9.12]
Non-mass enhancement (n = 62)
Type 1	27 (43.55)	12	44 [26–65]	0.91 [0.51–1.61]
Type 2	29 (46.77)	14	48 [29–68]	1.06 [0.63–1.81]
Type 3	6 (9.68)	3	50 [11.8–88.2]	1.14 [0.25–5.21]

	Beta
Predictor variable	coefficient	Odds ratio [95% CI]	P
Shape
Oval	0.00	1.00 (reference)	0.051
Round	4.35	77.66 [2.21–2,732.03]	0.017
Irregular	0.69	1.99 [0.30–13.03]	0.474
Margin
Circumscribed	0.00	1.00 (reference)	0.363
Irregular	3.00	20.14 [0.18–2,271.07]	0.21
Spiculated	3.89	48.98 [0.23–10,434.72]	0.16
Internal enhancement
Homogeneous	0.00	1.00 (reference)	0.039
Heterogeneous	3.33	27.87 [2.10–370.17]	0.012
Rim enhancement	3.54	34.39 [1.29–918.38]	0.035
Kinetic curve
Type 1	0.00	1.00 (reference)	< 0.001
Type 2	–1.74	0.18 [0.01–2.67]	0.211
Type 3	2.25	9.47 [1.27–70.35]	0.028

Publicação do Colégio Brasileiro de Radiologia e Diagnóstico por Imagem Av. Paulista, 37 - 7º andar - conjunto 71, 01311-902 - São Paulo - SP, Tel.: +55 11 3372-4541, Fax: 3285-1690, Fax: +55 11 3285-1690 - São Paulo - SP - Brazil
E-mail: radiologiabrasileira@cbr.org.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] Mailing address: Dr. João Ricardo Maltez Almeida. Clínica de Assistência à Mulher. Avenida ACM, 237, Ed. Prof. Carlos Aristides Maltez, Itaigara. Salvador, BA, Brazil, 41825-000. E-mail: jrmaltez.a@gmail.com.

Brasil

Brasil

Predictive performance of BI-RADS magnetic resonance imaging descriptors in the context of suspicious (category 4) findings* * Study conducted at the Clínica de Assistência à Mulher (CAM), Salvador, BA, Brazil.

Desempenho preditivo dos descritores de ressonância magnética do BI-RADS no contexto de achados suspeitos (categoria 4)

Abstract

Objective:

Materials and Methods:

Results:

Conclusion:

Predictive performance of BI-RADS magnetic resonance imaging descriptors in the context of suspicious (category 4) findings^* * Study conducted at the Clínica de Assistência à Mulher (CAM), Salvador, BA, Brazil.