Comparative outcomes of prepectoral subpectoral robot-assisted immediate prosthetic reconstruction following nipple-sparing mastectomy.

Introduction
Since Toesca et al. first reported robot-assisted nipple-sparing mastectomy (RANSM) using a lateral axillary incision in 2015 (1), this approach has been progressively integrated with robot-assisted immediate prosthetic breast reconstruction across multiple clinical settings (2,3). The Da Vinci (Intuitive, Sunnyvale, CA, USA) platform’s three-dimensional (3D) imaging capabilities and instrument flexibility enable complex surgical procedures through small incisions, which have been associated with improved patient satisfaction and favorable short-term outcomes compared with conventional methods, although current evidence remains largely observational (4). Additionally, the complication rate, including wound healing and nipple necrosis, has been reported to be low, and oncologic safety has been suggested but not yet fully established (3,5).
Prepectoral reconstruction demonstrates superiority over subpectoral reconstruction in implant-based breast reconstruction, offering simplified implant placement and fewer complications such as capsular contracture and animation deformity (6). Several reports also support its application in robot-assisted immediate prosthetic breast reconstruction (7,8). However, subpectoral reconstruction remains valuable, requiring less acellular dermal matrix (ADM) and providing a natural upper pole contour (9). Thus, comparing outcomes, esthetics, and complications in robotic surgery is essential.
Since 2021, our team has performed RANSM, followed by robot-assisted immediate prosthetic breast reconstruction using the Da Vinci X system. This case series presents our experience, comparing surgical outcomes and breast shape between prepectoral and subpectoral pockets. We hypothesized that the prepectoral approach in robot-assisted reconstruction would demonstrate superior esthetic outcomes and comparable safety compared with the subpectoral approach. We also discuss complications related to thoracic cage contour in subpectoral robotic reconstruction. We present this article in accordance with the STROBE reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-405/rc).

Methods

Patient selection
This single-institute retrospective analysis at Incheon St. Mary’s Hospital, College of Medicine, The Catholic University of Korea reviewed data of patients with breast cancer who underwent RANSM with immediate prosthetic reconstruction using the Da Vinci X system between October 2021 and May 2024. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Incheon St. Mary’s Hospital, College of Medicine, The Catholic University of Korea (No. OC25RISI0093). Written informed consent was obtained from all participants prior to their inclusion in the study. Both unilateral and bilateral cases were reviewed, using either one-stage permanent implants or two-stage expander-to-implant methods.
All patients diagnosed via core needle biopsy underwent standard preoperative imaging [ultrasonography, mammography, magnetic resonance imaging, chest/abdominopelvic computed tomography (CT), and bone scan]. A single general surgeon and a plastic surgeon (J.Y.L.) performed all procedures. Except for one case, all reconstructions used 150–500 cc Mentor® MemoryGel™ (Mentor Corp., Santa Barbara, CA, USA) implants placed in prepectoral or subpectoral pockets. One case used a Mentor® CPX™ 4 expander for two-stage reconstruction. ADM (CGDerm™ One-Step; CGBIO, Seoul, Republic of Korea or MegaDerm® HD; L&C BIO, Seoul, Republic of Korea) was chosen based on implant size, plane, and required thickness.
Demographic information, including age, body mass index (BMI), comorbidities, smoking status, biopsy results, and diagnosis, was collected for both groups. Surgical data included mastectomy type, axillary procedures, specimen weight, and use of chemotherapy/radiation. Implant type, ADM profile, and insertion plane were also documented.
Patients were allocated to either the prepectoral or subpectoral group based on the availability of square or rectangular ADMs, pectoralis major (PM) integrity, thoracic contour, and skin-flap thickness. Patient preference after preoperative counseling and intraoperative assessment by the senior surgeon were also considered.
Analyses were primarily conducted at the per-breast level, with each reconstructed breast regarded as an independent unit of analysis, allowing for detailed assessment of surgical outcomes while accounting for potential within-patient variability. This study evaluated outcomes exclusively within robot-assisted cases and did not aim to isolate the independent effects of the robotic approach or implant plane.
Surgical outcomes included operation time (measured from the start of surgery by the general surgeon to the completion of reconstruction by the plastic surgeon), console time (time spent by the plastic surgeon operating the console after docking), blood loss (difference between preoperative hemoglobin and hemoglobin levels on the 2nd–3rd postoperative day), hospital stay, and complications. Esthetic results were assessed at ≥1 year using clinical photographs and the Harvard/National Surgical Adjuvant Breast and Bowel Project (NSABP) scale (1= excellent to 4= poor), rated by seven independent plastic surgeons. All seven independent evaluators were blinded to the implant plane and assessed photographs without information regarding reconstruction type to minimize bias. Before evaluation, they participated in an anchoring session using representative sample images to standardize scoring criteria.

Surgical technique
The preoperative design included marking the midline, midclavicular point, anterior axillary line, and inframammary fold (IMF) to map breast contours. The upper breast pole was defined by palpating the breast-rib boundary. Implant selection involved measuring nipple positions relative to anatomical landmarks and assessing breast mound dimensions, including width, height, projection, and nipple-to-IMF distance, symmetrically on both sides.
Under general anesthesia with arms abducted, RANSM was performed via a 4–5-cm vertical anterior axillary incision. Following mastectomy, specimen size and weight, along with preoperative data, guided implant selection. After patient handoff, the surgical field was sterilized, and robot-assisted prosthetic reconstruction was initiated. Indocyanine green angiography was used to assess skin and nipple-areolar complex perfusion.
For prepectoral pocket creation, IMF tacking sutures were placed through the axillary incision. With the patient positioned upright, implant size was selected based on preoperative data, mastectomy specimen characteristics, and skin tension. A sizer was inserted to confirm the contour. Slits were created in the ADM, which was wrapped around the implant sizer and sutured to form a pocket, leaving the lateral side open for future implant insertion. After removing the sizer, the pocket-shaped ADM was folded, inserted through the axillary incision, and unfolded in place with the correct orientation. Following the Da Vinci X robot docking, the ADM was secured to the PM superomedially and to the IMF inferiorly using #2-0 Vicryl sutures, with tacking sutures placed along the chest wall to form the pocket.
For subpectoral pocket creation, the inferolateral border of the PM was dissected through the axillary incision using electrocautery and LigaSure (Stryker, Portage, MI, USA), elevating the muscle flap superomedially. The ADM was slit and trimmed to cover the implant and pre-inserted into the breast pocket. After the Da Vinci X platform was docked, the inferomedial border of the PM was dissected further. The ADM was secured to the muscle and IMF line, forming the subpectoral pocket. Following undocking, saline irrigation and bleeding control were performed. An implant sizer was inserted, and the patient was seated upright to assess breast symmetry and select the implant.
In both methods, the surgical assistant marked the breast pocket boundaries with a needle, guiding the console operator for precise dissection and suture placement. After re-draping, two Jackson-Pratt drains were positioned in the superior and inferior pocket areas. Antibiotic irrigation was performed, and the permanent implant was inserted. The incision was closed in anatomical layers, and a compressive dressing was applied using a surgical bra.

Postoperative management
Postoperatively, topical nitric oxide (RECTOGESIC®; Care Pharmaceuticals, New South Wales, Australia) was applied twice daily, and 2 L of oxygen were provided to maintain saturation. The drain was removed once output was under 20 cc, and the patient was discharged.
A compression bra was worn for 3 months and a breast belt for 1 month. Follow-up visits were scheduled at 2, 3, 6, and 12 months and annually thereafter. Massage, laser therapy, and scar gel application were encouraged. Clinical photographs were taken at 6 months and 1 year for breast assessment.

Statistical analyses
Statistical analyses were performed using GraphPad Prism 9.0.1 (GraphPad Software Inc., San Diego, CA, USA). Continuous variables were analyzed using the Mann-Whitney U test, and categorical variables were analyzed using the chi-square test (or Fisher’s exact test when expected cell counts were <5). Inter-rater reliability was assessed using the intraclass correlation coefficient (ICC) based on a two-way mixed-effects model for consistency agreement (ICC[3, 1]) using STATA (StataNow 19.5 BE, StataCorp, College Station, TX, USA). Statistical significance was set at P<0.05.

Results
Between October 2021 and May 2024, 10 patients underwent RANSM with immediate prosthetic reconstruction using the Da Vinci X system. Five patients received prepectoral implants and five received subpectoral implants. One subpectoral case required revision, resulting in a total of 11 surgeries and 17 reconstructed breasts (five unilateral and six bilateral cases), including the revision. Analyses were primarily performed at the per-breast level. Results remained internally consistent, supporting the validity of this analytical approach.
The median age was 45.5 years (P=0.50), and BMI was 22.4 kg/m2 (P=0.17), with 20% patients being former smokers (P=0.44); no significant differences were observed between the groups. Excluding the single revision procedure performed solely by the plastic surgery team, 16 mastectomies were performed, including 15 therapeutic procedures and 1 prophylactic mastectomy. Among the 15 cancer-related cases, six patients had invasive ductal carcinoma (IDC), five had ductal carcinoma in situ (DCIS), three had atypical ductal hyperplasia, and one had papillary neoplasm. Final diagnoses revealed seven cases of IDC, six of DCIS, and two with no residual tumor (post-mammotome or neoadjuvant chemotherapy).
All 16 mastectomies were performed using the nipple-sparing technique. The median specimen weight was 235.15 g (P=0.17). Sentinel node biopsy was performed in 10 breasts, and axillary dissection in three. Five patients did not undergo any axillary surgery. Postoperatively, two patients received radiation therapy (P=0.44), and three received chemotherapies, including one after neoadjuvant treatment (P=0.17), with no significant group differences (Table 1). Other potential confounders, including PM integrity, ADM thickness and type, prior radiation, and smoking status, were comparable between groups.
Implants were inserted in one stage for 15 breasts, with one two-stage case in the prepectoral group, using a tissue expander. All surgeries involved the use of ADM to form the implant pocket. ADM thickness ranged from 1.5–3.0 mm, except for one subpectoral case (1.0–1.5 mm, Table 2). Mean operation time was 313.4 min (P=0.60), console time 57.1 min (P=0.60), blood loss 2.28 g/dL (P=0.60), and hospital stay 12.4 days (P=0.22), with no significant group differences. The mean follow-up duration was 14 months, ranging from 12 to 26 months. The relatively long hospitalization period reflects institutional practice patterns rather than prolonged recovery. In the prepectoral group, one case of seroma resolved after aspiration, and one case of minor areolar necrosis (post-neoadjuvant chemotherapy) required debridement and closure. In the subpectoral group, use of a thin ADM (1.0–1.5 mm) led to bottoming out, requiring robotic revision. One patient who smoked developed recurrent cellulitis 11 months postoperatively, requiring implant removal. Animation deformity occurred in two patients, both involving PM defects, which were repaired using ADM patches (Table 3). An exploratory analysis was conducted to assess whether the difference between mastectomy specimen weight and implant volume correlated with postoperative complications (10). In our small-volume cohort, logistic regression revealed an opposite trend to previous reports (OR =0.98, P=0.21), indicating that implant oversizing rather than underfilling may increase complication risk in smaller reconstructions.
Esthetic evaluation was performed by seven plastic surgeons using the Harvard/NSABP scale. The mean scores of the prepectoral and subpectoral groups were 2.42 and 3.65, respectively (P=0.04), showing significantly better results in the prepectoral group (Table 4). A two-way mixed-effects model for consistency (ICC[3,
1]) showed excellent inter-rater reliability [single-measure ICC =0.77, 95% confidence interval (CI): 0.57–0.92; average-measure ICC =0.96, 95% CI: 0.90–0.99, P<0.001], indicating highly consistent scoring among raters.
To enhance clinical interpretation, three representative cases were deliberately selected to illustrate distinct clinical scenarios: (I) an uncomplicated prepectoral reconstruction demonstrating typical aesthetic and functional outcomes; (II) a subpectoral reconstruction complicated by partial PM injury; and (III) a subpectoral reconstruction affected by thoracic contour leading to animation deformity. These cases were chosen to exemplify the key technical and anatomical factors discussed in this study.

Case 1—prepectoral group
A 45-year-old woman presented with bilateral breast masses detected on ultrasound screening (Figure 1A). Core needle biopsy showed DCIS in the right breast and atypical ductal hyperplasia in the left breast. Bilateral nipple-sparing mastectomy with immediate reconstruction was planned. To avoid visible scars, the patient opted for robot-assisted surgery.
The general surgery team performed bilateral nipple-sparing mastectomy with right sentinel node biopsy via a 4.5-cm axillary incision. The patient was then transferred to the plastic surgery team. Both incisions exposed the PM without any damage.
Using the described method, the plastic surgery team created a prepectoral pocket with ADM (CGDerm™ One-Step, 16×16 cm, thickness 2–3 mm, CGBIO) and performed reconstruction with a 190 cc implant (Mentor® MemoryGel™ Smooth Round Moderate Classic Profile, Mentor Corp.).
Final pathology confirmed DCIS in both breasts. The patient did not require adjuvant therapy and underwent routine follow-up. There were no complications, the scars were stable, and the breasts were symmetrical (Figure 1B and Figure S1A,S1B).

Case 2—subpectoral group 1
A 42-year-old female presented with multiple masses in both breasts detected on ultrasound screening (Figure 2A). A core needle biopsy confirmed IDC in the left breast and atypical ductal hyperplasia with microcalcifications in the right breast. With consent, the general surgery team planned bilateral RANSM and referred her for immediate prosthetic breast reconstruction.
The general surgery team performed bilateral RANSM via a 4-cm extramammary axillary incision and sentinel lymph node biopsy of the left breast. Subsequently, the patient was transferred to the plastic surgery team. Axillary incisions exposed the PM, with a partial defect on the inferomedial side of the right PM (Figure 2B, arrowheads).
Due to inward indentation of the thoracic wall, visual inspection and chest CT revealed that accessing the medial side through the axillary port was difficult. The plastic surgery team created a subpectoral pocket with ADM (MegaDerm® HD 6×16 cm, 8×16 cm, thickness 1.5–2.3 mm, L&C BIO) and used an ADM patch to cover the right PM defect (Figure 2C, arrow). Reconstruction was completed with 255 cc breast implants (Mentor® MemoryGel™ Smooth Round Moderate Classic Profile, Mentor Corp.).
Final pathology confirmed IDC in the left breast and DCIS in the right breast. No adjuvant therapy was required, and both teams followed routine postoperative schedules. At the 1-year follow-up, an animation deformity was noted in the superolateral area of the right breast, where the ADM patch was used, because of PM damage (Figure 2D,2E). No deformities were observed in the left breast, with normal PM coverage. Scars in both axillae were stable (Figure S1C,S1D).

Case 3—subpectoral group 2
A 50-year-old female patient was referred for general surgery after an enhancing lesion was detected in the left breast on chest CT during a gynecological follow-up (Figure 3A). Breast ultrasonography revealed masses in both breasts, and core biopsy confirmed DCIS in the left breast and a papillary neoplasm in the right breast.
With consent, the general surgery team performed bilateral RANSM through a 4-cm extramammary axillary incision and sentinel lymph node biopsy of the left breast. The patient was subsequently transferred to the plastic surgery team. Axillary incisions exposed the PM, revealing a defect along the lower border of both muscles.
In this case, visual inspection and chest CT showed an anteroposteriorly elongated thoracic cavity, limiting medial access via the axillary port. A subpectoral pocket was created using ADM (MegaDerm® HD and CGDerm™ One-Step, 8×16 cm, 2–3 mm thick). A strip-shaped lower PM defect was covered with an ADM patch (MegaDerm® HD 6×16 cm, 1.5–2.3 mm) (Figure 3B,3C). Reconstruction was completed with 320 cc implants (Mentor® MemoryGel™ Smooth Round Moderate Classic Profile) in both breasts.
Final pathology confirmed IDC in the left breast and DCIS in the right breast. Routine postoperative follow-up was performed, and no adjuvant therapy was needed. At the 1-year follow-up, animation deformity was observed in both breasts where ADM patches were used (Figure 3D,3E), with resting depression in the superolateral area and visible motion during muscle contraction due to PM damage and altered coverage vectors. Axillary scars remained stable (Figure S1E,S1F).

Discussion
In this study, we outlined our early experience with robot-assisted subpectoral direct-to-implant (DTI) immediate breast reconstruction (IBR) in RANSM, highlighting its potential. There is a growing shift from subpectoral to prepectoral DTI owing to easier implant placement and preserved PM function, which lowers animation deformity (6,11,12). Recent data from Campbell and Losken have further supported the safety and aesthetic advantages of the prepectoral approach (13). However, subpectoral DTI remains valuable in RANSM-IBR. Its advantages include better upper medial pole contour and reduced ADM use (9). However, in our series, ADM was applied in all cases, primarily due to the technical and anatomical characteristics of robotic reconstruction. ADM was used to reinforce pocket coverage, prevent implant migration, and minimize capsular contracture, particularly given the limited postoperative access inherent to robotic procedures. Even in subpectoral cases with an intact pectoralis muscle, ADM served to distribute pressure evenly and stabilize implant position.
The Da Vinci robotic platform is not currently approved by the U.S. Food and Drug administration (FDA) for breast surgery, although several clinical trials are ongoing to evaluate its safety and feasibility. Recent multicenter data from Farr et al. have also demonstrated encouraging short-term outcomes supporting the feasibility of robotic-assisted mastectomy (14). Regulatory approval status differs by region, and other robotic platforms are emerging to expand potential applications in the breast. This regulatory context highlights the experimental nature of robotic breast reconstruction and underscores the importance of ethical oversight.
The longer operative time in case 1 reflected the early learning curve for the prepectoral robotic approach. Subsequent cases showed markedly reduced durations as the surgical team gained experience. Although the mean operative time for the prepectoral group was slightly longer than that for the subpectoral group, this difference was not statistically significant and is expected to diminish with increased procedural familiarity.
Although subpectoral DTI in RANSM-IBR has been reported, most studies involved single cases or small cohorts with limited follow-up (15), lacking comprehensive long-term data. The predominant complication reported was animation deformity (16). In the present study, animation deformity was observed in association with the integrity of the PM muscle. Cases with preserved PM continuity tended to show mild, clinically insignificant deformity, whereas those with partial PM discontinuity demonstrated more noticeable deformity.
In case 2, minor disruption to the inner right PM required ADM reinforcement; however, a noticeable depression appeared during contraction. In Case 3, discontinuity along the lower border of the PM corresponded with a deeper depression, visible both at rest and during movement. These findings suggest that insufficient muscular coverage of the implant may contribute to upward muscle vector shifts during contraction, accentuating animation deformity. This likely stemmed from the patient’s thoracic shape and the straight-line approach of the robotic arm via a fixed lateral port (Figure 4A). To our knowledge, no prior studies have addressed robotic arm-blind spots related to the thoracic anatomy.
As demonstrated in Figure 4B, the robotic arm easily accesses the medial chest when the innermost area is elevated. However, in patients with inner chest wall depression, the arm must angle upwards to reach this area, increasing the tension on the port site skin (Figure 4C). Limited access may also induce some muscle removal during breast tissue dissection, risking PM damage. In cases with a large anteroposterior thoracic diameter and a lower lateral port (Figure 4D), dissection may follow a “shoveling” motion along the PM’s anterior surface, causing further injury. Taken together, these observations indicate that maintaining PM integrity is likely important for minimizing animation deformity and ensuring optimal outcomes.
One patient developed bottoming-out with mild contracture in the left breast 1 year after subpectoral implantation. She requested implant exchange to a larger implant, and although a textured implant was planned, a smooth implant was ultimately used owing to supply issues, which likely contributed to the complication. A thinner-than-intended ADM was also used because of stock limitations, and its absorption likely failed to adequately support the implant base. The patient underwent robot-assisted revision surgery, which enabled precise suturing and improved 3D visualization, offering clear benefits for complex revisions.
In our cases, no major complications were encountered. However, successful RANSM-IBR depends critically on effective skin flap management. Essential factors include appropriate skin flap elevation during mastectomy, limited CO2 insufflation, gentle robotic arm handling to reduce port tension, and careful port-margin revision. Postoperative perfusion can also be improved with nitroglycerin ointment to dilate vessels or supplemental oxygen therapy (17,18).
This study was limited by its small sample size and retrospective design, which may have introduced a selection bias and potential confounding from the combined analysis of robotic and implant plane variables. Other limitations include the absence of patient-reported outcome measures and a relatively short follow-up period. Moreover, the cost-effectiveness and learning-curve aspects of robotic reconstruction were beyond the scope of this study but represent important areas for future investigation. Despite these limitations, our initial experience with subpectoral DTI in RANSM-IBR provides preliminary insights that may help refine future indications and techniques for robotic implant-based breast reconstruction. Larger, prospective studies and randomized trials are warranted to more comprehensively assess comparative outcomes.

Conclusions
In conclusion, this study compared robot-assisted prepectoral and subpectoral DTI IBR following RANSM. The prepectoral approach generally yielded better esthetic outcomes, as indicated by a lower Harvard/NSABP score, and was associated with fewer complications, such as animation deformity, which is more commonly observed in the subpectoral group due to PM involvement. The subpectoral approach offers advantages such as a more natural upper breast pole and reduced ADM use; however, it carries a higher risk of animation deformity, especially with PM damage. These findings highlight the need for careful PM management to reduce complications and ensure optimal outcomes. Our early experience suggests that prepectoral robot-assisted reconstruction tends to offer promising esthetic outcomes and fewer complications compared with the subpectoral approach. However, subpectoral DTI remains a viable option in selected cases. Further research with larger samples and longer follow-up is warranted to validate these preliminary findings and refine robotic-assisted techniques for improved precision and safety.

Supplementary
The article’s supplementary files as