Accuracy of 16/18G core needle biopsy for ultrasound-visible breast lesions

Percutaneous ultrasound- and stereotactic-guided breast core needle biopsy (CNB) is widely used as a reliable alternative to surgical biopsy to obtain a histological diagnosis for imagery-visible suspicious breast lesions [1‐3]. Stereotactic-guided percutaneous breast biopsy is mostly used for micro-calcifications, while ultrasound-guided biopsy is mostly used for masses and architectural distortions. Previous studies showed that stereotactic-guided CNB using a large needle gauge provided an accurate diagnosis [4, 5].

However, the accuracy of using 16G and 18G needles for ultrasound-guided CNB, especially for ultrasound-visible mass lesions, have not been fully assessed. Few distinctions are made between stereotactic and ultrasound-guidance CNB when needle size is discussed [6]. Little information is currently available about needle size selection for ultrasound-guided biopsies of lesions with different imagery features, and most of this information comes from non-breast lesions [7], or from studies that used 14G needles only [8‐18]. Therefore, one may raise the question about whether ultrasound-guided CNB with smaller needles, such as 16G and 18G, can have diagnostic value for ultrasound-visible lesions, and if there is any difference in the accuracy of biopsies of breast lesions with different imaging features.

In this article, we present a large series of breast lesions that underwent ultrasound-guided 16G or 18G CNB followed by subsequent surgical excision. We hypothesized that smaller gauge CNB provides results that are usable in a routine clinical setting. The objectives were to assess the accuracy of this technique for ultrasound-visible breast lesions, and to analyze the effects of lesion features. The results of this study could allow the use of needles that are less expensive and less traumatic for the patient, without impairing the accuracy of the biopsy.

Patients’ mean age was 48.9 ± 11.4 years (range: 15–85) for 16G CNB and 49.8 ± 11.3 years (range: 16–87) for 18G CNB. Among all lesions, 885 lesions were BI-RADS 4 or 5, and 70 were BI-RADS 3. The mean lesion size was 22 ± 12 mm (range: 3–100) in the 16G group and 23 ± 13 mm (range: 4–83) in the 18G group. The mean number of cores for 16G and 18G CNB were 2.8 ± 0.5 (range: 2–6) and 2.8 ± 0.5 (range 2–6), respectively. There were no differences in the breast lesion characteristics between the two gauges (Table 1).

Pathological examination of the CNB revealed that malignant lesions accounted for 84.1% (n = 803) of CNB, high-risk lesions accounted for 8.4% (n = 80), and benign lesions accounted for 7.5% (n = 72) (Table 2).

False-negative rates, sensitivities, underestimation rates, and agreement rates of 16G and 18G CNB compared with the final surgical histological results were calculated according to the CNB gauge, lesion type, lesion size, and calcifications (Table 3). False-negative rates were 1.4% for 16G and 18G CNB. Agreement rates between histological findings of CNB and surgery were 92.4% for 16G and 92.8% for 18G CNB, and there was no difference between the two gauges (P >0.01). However, there was a risk of underestimation with 16G and 18G CNB. The high-risk underestimation rates were 24/50 (48.0%) for 16G CNB and 16/30 (53.3%) for 18G CNB. The DCIS underestimation rates were 12/26 (46.2%) for 16G CNB and 7/17 (41.2%) for 18G CNB. There were no statistical differences between the two gauges.

For both 16G and 18G CNB, the absolute sensitivities and agreement rates were better for mass lesions than for non-mass lesions, reaching statistical significance in 18G CNB (absolute sensitivity: mass: 94.2% vs. non-mass: 89.5%; agreement: mass: 93.7% vs. non-mass: 78.3%; all P <0.01). Results showed a significant trend toward a better agreement of CNB with increasing mass size. Indeed, false-negative rates of CNB were the highest and agreement rates were the lowest when mass size was ≤10 mm (accounting for 13.9% of CNB in the present study). In addition, although lesions with calcification had higher agreement rates (18G: 94.4 vs. 91.9%; 16G: 95.5 vs. 91.0%) and lower underestimation rates (18G: 40.0 vs. 53.0%; 16G: 33.3 vs. 51.7%) than lesions without calcification (Table 3), there were no significant differences between the two groups.

Minimally invasive percutaneous CNB is used for ultrasound-visible breast lesions, but no study has previously assessed the accuracy of CNB using 16G or 18G needles. The objective was then to assess the accuracy of this technique for ultrasound-visible breast lesions, and to analyze the effects of lesions’ image features. Results showed that false-negative rates were low for 16G and 18G CNB, and that agreement rates between histological findings of CNB and surgery were >92% for 16G and 18G CNB. Agreements were better for mass lesions than for non-mass lesions. Agreement rates were lower for mass lesions with a size ≤10 mm.

Ultrasound-guided CNB has become the first biopsy choice for ultrasound-visible lesions [21]. It offers several advantages over stereotactic guidance, including no radiation, full real-time control of the needle position, and patients’ comfort without breast compression. It is also more flexible for biopsy of multiple lesions or in difficult places (such as the axilla or near the nipple) [22]. Previous studies showed that ultrasound-guided CNB had a high accuracy when using 14G needles [2, 10, 13, 16, 20‐22]. However, very little information is currently available about how these results may vary when the biopsies are performed using smaller size needles, and most of these results are from non-breast studies [7].

China has different ethnic populations and different medical treatment situations compared with most Western countries. In addition, Asian women are observed to have smaller breasts and denser mammary parenchyma than Caucasian women [23]. Due to the magnitude of breast cancer risk factors, genetic and/or environmental discrepancies, breast cancer in China shows aggressive behaviors. In patients with a large tumor size (>2–5 cm, accounting for 51.3% of cases), the mean age is in the 40s, which is a decade earlier than what is reported in Western countries [24, 25]. Ultrasound-guided biopsies are more sensitive to detect breast cancers than stereotactic-guided ones in young women, especially for invasive carcinoma [26]. Further, the use of ultrasound in a screening setting is widespread in China. Ultrasound-guided CNB is the first choice for performing a percutaneous biopsy for most lesions seen or found to be highly suspicious on ultrasound. In China, 16G or 18G core needles are commonly used in most medical institutes. This is because in regards to mass lesions, target accuracy and proper procedures are essential for predicting the final results, regardless of the size of the core needle or the number of core samples [6]. Moreover, the smaller the needle diameter, the lower the friction will be with the surrounding tissues. Therefore, the puncture is easier to make and the strength of the shot is greater, which is helpful when the breast tissue is dense and causes difficulties in positioning the needle when the breast is relatively mobile. In addition, at least in theory, using small needles should decrease bleeding and pain [22]. The selection of a small gauge core needle may be reasonable from the viewpoint of needle handling, and the samples harvested by ultrasound-guided 16G or 18G CNB are approved by most pathologists and surgeons in our hospital. Then, if we assume that tissue samples obtained by 16G or 18G CNB could be adequate for an accurate diagnosis, why not use smaller needle for easier biopsy procedures and fewer complications?

In the present study, both 16G and 18G CNB had overall false-negative rates of 1.4%, which is acceptable compared with previously reported rates of 0% to 12% [13, 16, 20]. These results confirmed that ultrasound-guided 16G and 18G CNB are accurate and reliable alternatives to surgical biopsy for ultrasound-visible breast lesions.

However, a significant problem with CNB is the high rate of histological underestimation. A meta-analysis of underestimation of high-risk lesions in stereotactic-guided biopsy showed a rate of 40% [27]. Studies demonstrated that among lesions yielding DCIS at ultrasound-guided CNB, surgeries revealed infiltrating carcinoma in 16% to 55.5% of cases [21]. The rate of ADH underestimation with ultrasound-guided 14G CNB was even more highly variable (0% to 100%, mean 53%) [28, 29]. Similar to these previous studies, the high-risk underestimation and DCIS underestimation rates of 16G and 18G CNB were also relatively high in the present study. Therefore, we recommend vacuum-assisted CNB or complete surgical excision of all such lesions detected by CNB, to ensure the detection of any coexisting malignancy or invasive cancer.

With the improvement of ultrasound resolution and contrast, microcalcifications contrasting with background hypoechoic areas or duct-like structures can now be detected by ultrasound [30]. It is very important to classify non-mass lesions visible using ultrasound because these lesions are more histologically heterogeneous than mass lesions, and usually include more DCIS and non-palpable lesions [6]. In the present study, there was a significant difference in the agreement between mass and non-mass lesions for 18G CNB. This indicates that for smaller caliber CNB performed on non-mass lesions, especially 18G, the diagnosis value is lower than biopsies using a larger core needle or vacuum-assisted CNB [31].

Small tumor size is proven to contribute to the inaccuracy of CNB [8]. In our study, the agreements decreased with smaller lesions. This finding is similar to a previous study performed using 14G CNB [8]. This may be the result of the poor lesion or needle visualization within the small lesion, influenced by partial volume effect artifacts towards the periphery of the lesion. Therefore, radiologists must be aware of these difficulties and should pay extra attention in lesions ≤10 mm when performing ultrasound-guided CNB. Larger caliber CNB or vacuum-assisted CNB is recommended.

There are certain limitations in the present study. First, lesions with benign results at CNB without subsequent surgical excision have been excluded. Therefore, a selection bias may exist, and it is possible that there were more false-negative diagnoses in the excluded cases. That is, the true sensitivity of CNB using 16G or 18G needles is equal to or lower than that calculated in this study. Maybe this is the reason why our false-negative rates are relative high compared with previous results (Table 4). Although the reports in Table 4 showed there were few missed malignant cases in the follow-up period (0 to 3 cases) [11‐14], further study to follow patients with a benign CNB result who did not undergo excisional biopsy is necessary to know the actual sensitivities. Secondly, not every case was harvested with the same number of core samples, although a previous study indicated that two cores are sufficient to diagnose breast cancer, assuming that no technical error occurred [15]. Finally, this study did not make any direct comparison with 14G ultrasound-guided CNB. However, our data are compatible with data obtained using 14G CNB in previous studies [8‐18].

In conclusion, for ultrasound visible breast lesions, 16G and 18G ultrasound-guided CNB can be used as an accurate diagnostic alternative to surgical biopsy, at least in China. For mass lesions visible on ultrasound with a size >10 mm, 18G and 16G CNB are accurate and should be the first choice for ultrasound-guided CNB of these lesions. For mass lesions with a size ≤10 mm or for non-mass lesions, large caliber CNB or vacuum-assisted CNB is necessary. Considering the high rate of underestimation of high risk lesions by ultrasound-guided CNB, large volume vacuum-assisted biopsy or surgical excision is strongly recommended.