Is there a correlation between signal intensity ratio of 3T MRI and molecular subtypes of breast cancer?

Yuanping Tao; Hongyan Chao; Yue Ren; Wanjun Zheng; Yuyou Chen; Liang Du; Huanguo Li; Feng Cui

doi:10.2478/raon-2026-0029

.blurhash-client-img { display: none !important; }

Is there a correlation between signal intensity ratio of 3T MRI and molecular subtypes of breast cancer?

Radiology and Oncology

AHEAD OF PRINT

By: Yuanping Tao, Hongyan Chao, Yue Ren, Wanjun Zheng, Yuyou Chen, Liang Du, Huanguo Li and Feng Cui

Open Access

|Jul 2026

Figures & Tables

Inclusion and exclusion criteria and the data collection flowchart showing the study population.

MR images of a 48 years-old patient diagnosed with luminal B subtype breast cancer. T1 weight image **(A)**, T2 fatsuppression sequence image **(B)**, DWI image **(C)** and ADC map **(D)** was shown. T1 and T2 signal intensity ratio (A and B) were calculated by proportioning the average signal intensity of the lesion (red circle) to the average signal intensity of the pectoral muscle (yellow circle). Red arrows show the lesions and yellow arrows show the pectoral muscles.
ADC = apparent diffusion coefficient; DWI = diffusion-weighted imaging

TABLE 1.

Clinical characteristics of patients and MRI morphological features for different molecular subtypes

Parameters	Luminal A (n = 52)	Luminal B (n = 135)	HER2-enriched (n = 31)	TN (n = 45)	p
ALNM					0.010
Yes	9 (17.3)	58 (43.0)	12 (38.7)	14 (31.1)
No	43 (82.7)	77 (57.0)	19 (61.3)	31 (68.9)
History of hormone use					0.038
Yes	3 (5.8)	2 (1.5)	3 (9.7)	0 (0.0)
No	49 (94.2)	133 (98.5)	28 (90.3)	45 (100.0)
Tumor morphology					0.053
Mass	43 (82.7)	114 (84.4)	20 (64.5)	39 (86.7)
Non-mass	9 (17.3)	21 (15.6)	11 (35.5)	6 (13.3)
Margin					0.206
Circumscribed	21 (40.4)	39 (28.9)	13 (41.9)	19 (42.2)
Not-circumscribed	31 (59.6)	96 (71.1)	18 (58.1)	26 (57.8)
Dynamic curve curve					0.248
type I	8 (15.4)	13 (9.6)	4 (12.9)	7 (15.6)
type II	9 (17.3)	30 (22.2)	12 (38.7)	8 (17.8)
type III	35 (67.3)	92 (68.1)	15 (48.4)	30 (66.7)

1 Categorical variables are numbers with percentages in parentheses. Bolded text indicates statistical significance (p < 0.05).

1 ALNM = axillary lymph node metastasis; HER2 = human epidermal growth factor receptor 2; TN = triple-negative

TABLE 2.

Intraclass correlation coefficient (ICC) analysis between the two readers

Parameters	ICC (95%CI)
T1 SIR	0.970 (0.961–0.976)
T2 SIR	0.967 (0.957–0.974)
Mean ADC (×10⁻³ mm²/s)	0.898 (0.871–0.919)
Tumor diameter (cm)	0.923 (0.900–0.940)

1 ADC = apparent diffusion coefficient; T1 SIR = T1 signal intensity ratio; T2 SIR = T2 signal intensity ratio

TABLE 3.

MRI quantitative parameters for molecular subtypes of breast cancer

Parameters	Luminal A (n = 52)	Luminal B (n = 135)	HER2-enriched (n = 31)	TNBC (n = 45)	p
T1 SIR	1.318 (1.179–1.630)	1.366 (1.209–1.620)	1.515 (1.268–1.788)	1.254 (1.141–1.401)	0.026
T2 SIR	3.137±1.156	3.163±1.186	3.280±1.289	3.689±1.093	0.009
Mean ADC (×10⁻³ mm²/s)	1.037±0.291	0.912±0.211	1.102±0.297	0.900±0.164	0.000
Tumor diameter (cm)	1.200 (0.887–1.762)	1.900 (1.325–2.725)	1.150 (0.600–2.750)	2.550 (1.700–3.750)	0.000

1 Normally distributed variables are expressed as the mean ± SD, and abnormally distributed variables are expressed as the median with the first and third interquartile range (IQR). Bolded text indicates statistical significance (p < 0.05).

1 ADC = apparent diffusion coefficient; HER2 = human epidermal growth factor receptor 2; T1 SIR = T1 signal intensity ratio; T2 SIR = T2 signal intensity ratio; TN = triple-negative

TABLE 4.

ROC analysis of MRI quantitative parameters for separating one subtype from all other subtypes

Parameters	AUC	Cutoff value	Sensitivity (%)	Specificity (%)	p*
Luminal A vs. all others
T1 SIR	0.506	≤ 1.270	46.15 (40.1, 52.2)	61.61 (55.7, 67.5)	0.640
T2 SIR	0.550	≤ 3.000	59.62 (53.7, 65.6)	54.50 (48.5, 60.5)	0.397
Mean ADC (×10⁻³ mm²/s)	0.612	> 0.980	46.15 (40.3, 52.0)	70.62 (65.3, 76.0)	0.022
Tumor diameter(cm)	0.768	≤l.900	80.00 (73.1, 86.9)	64.2 (57.0, 71.4)	0.016
Luminal B vs. all others
T1 SIR	0.510	>1.291	62.96 (56.2, 69.7)	48.44 (41.7, 55.2)	0.181
T2 SIR	0.563	≤ 3.298	65.93 (59.6, 72.3)	45.31 (38.6, 52.0)	0.169
Mean ADC (×10⁻³ mm²/s)	0.624	≤ 0.978	77.78 (71.6, 84.0)	44.53 (37.8, 51.3)	0.012
Tumor diameter(cm)	0.540	>1.550	78.63 (71.7, 85.6)	35.58 (28.2, 42.9)	0.829
HER2-enriched vs. all others
T1 SIR	0.631	> 1.503	58.06 (50.3, 65.8)	69.40 (61.9, 76.9)	0.000
T2 SIR	0.517	> 2.815	67.74 (60.0, 75.5)	44.40 (36.6, 52.2)	0.927
Mean ADC (×10⁻³ mm²/s)	0.713	> 0.903	87.10 (81.5, 92.7)	53.02 (45.1, 60.9)	0.002
Tumor diameter(cm)	0.543	> 2.190	60.00 (52.4, 67.6)	60.20 (52.6, 67.8)	0.697
TN vs. all others
T1 SIR	0.608	≤ 1.286	64.44 (58.6, 70.2)	62.84 (57.0, 68.7)	0.176
T2 SIR	0.655	>3.157	68.89 (61.8, 76.0)	59.63 (51.9, 67.4)	0.020
Mean ADC (×10⁻³ mm²/s)	0.562	≤ 0.869	51.11 (44.1, 58.1)	62.39 (54.9, 69.9)	0.087
Tumor diameter(cm)	0.706	>2.050	74.36(65.3, 83.4)	57.69 (49.3, 66.1)	0.040

1* Corrected p values by Benjamini-Hochberg method. Data are values with 95% confidence interval in parentheses. Bolded text indicates statistical significance (p < 0.05).

1 ADC = apparent diffusion coefficient; AUC = area under the ROC curve; HER2 = human epidermal growth factor receptor 2; ROC = receiver operating characteristic; T1 SIR = T1 signal intensity ratio; T2 SIR = T2 signal intensity ratio; TN = triple-negative

The comparison of receiver operating characteristic (ROC) curve of single and combined parameters for distinguishing one subtype from all other subtypes.
ADC = apparent diffusion coefficient; ALNM = axillary lymph node metastasis; TN: triple-negative; T1 SIR = T1 signal intensity ratio; T2 SIR = T2 signal intensity ratio

TABLE 5.

ROC analysis of combined parameters for separating one subtype from all other subtypes

	AUC	Sensitivity (%)	Specificity (%)	p
Luminal A vs. all others
Mean ADC+ tumor diameter	0.781	82.22 (76.57, 87.87)	74.43 (68.99, 79.87)	< 0.001
Luminal B vs. all others
Mean ADC+ALNM	0.639	62.96 (57.12, 68.80)	60.94 (55.04, 66.84)	< 0.001
HER2-enriched vs. all others
T1 SIR + mean ADC+ non-mass	0.728	38.71(21.56, 55.86)	94.4(91.44, 97.36)	< 0.001
TN vs. all others
T2SIR+ tumor diameter	0.724	89.74 (85.64, 93.84)	52.20 (45.68, 58.72)	0.002

1 Bolded text indicates statistical significance (p < 0.05). Data are values with 95% confidence interval in parentheses.

1 ADC = apparent diffusion coefficient; ALNM = axillary lymph node metastasis; AUC = area under the ROC curve; HER2 = human epidermal growth factor receptor 2; TN = triple-negative; T1 SIR = T1 signal intensity ratio; T2 SIR = T2 signal intensity ratio

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.2478/raon-2026-0029 | Journal eISSN: 1581-3207 | Journal ISSN: 1318-2099

Journal RSS Feed

Language: English

Submitted on: Jan 23, 2026

Accepted on: Apr 11, 2026

Published on: Jul 4, 2026

Published by: Association of Radiology and Oncology

In partnership with: Paradigm Publishing Services

Publication frequency: 4 issues per year

Keywords:

breast cancer,

signal intensity ratio,

Related subjects:

© 2026 Yuanping Tao, Hongyan Chao, Yue Ren, Wanjun Zheng, Yuyou Chen, Liang Du, Huanguo Li, Feng Cui, published by Association of Radiology and Oncology
This work is licensed under the Creative Commons Attribution 4.0 License.

AHEAD OF PRINT

Is there a correlation between signal intensity ratio of 3T MRI and molecular subtypes of breast cancer?

Figures & Tables

FIGURE 1.

FIGURE 2.

TABLE 1.

TABLE 2.

TABLE 3.

TABLE 4.

FIGURE 3.

TABLE 5.

Paradigm

My account