Meme kanserinde neoadjuvan kemoterapiye yanıtı değerlendirmede dinamik meme MRG parametrelerinin doğruluk analizi
Yükleniyor...
Tarih
2021
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Ege Üniversitesi, Tıp Fakültesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Amaç: Meme kanserli olgularda neoadjuvan kemoterapiye (NAKT’ye) yanıtı değerlendirmede
dinamik meme manyetik rezonans görüntüleme (MRG) parametrelerinin, histopatolojik
inceleme verileri referans test kabul edilerek, yanıtı öngörmedeki performansı, değerlendirmede
etki eden morfolojik, histopatolojik ve moleküler özellikleri test etmektir.
Gereç-Yöntem: Bu retrospektif çalışmada, Ocak 2011-Şubat 2020 tarihleri arasında
hastanemizde NAKT adayı olan, tedavi öncesi ve sonrası (cerrahi öncesi) dinamik MRG
uygulanan, 3’ünde bilateral olmak üzere toplam 90 meme kanserli (ortalama yaş 46.03 ±9.58;
yaş aralığı 24-66 ) kadın ve 93 invaziv lezyon çalışma grubunu oluşturmuştur. Radyolojik tam
yanıt primer tümör bölgesinde erken dinamik fazda kontrast etkileşimi olmaması olarak
değerlendirilirken, patolojik tam yanıt in situ kanser (DKİS) varlığına bakılmaksızın meme
parankiminde invaziv-mikroinvaziv tümör yokluğu olarak kabul edilmiştir. Aksiller tutulum
radyopatolojik değerlendirmeye dahil edilmedi. NAKT’ye yanıtı tahmin etmede dinamik
MRG’nin tanısal performansı ( doğruluk, sensitivite, spesifite, negatif prediktivite (NPV), pozitif
prediktivite (PPV)), bunu etkileyen faktörler ( kitle, boyutu, hacmi, tipi, meme parankim
kontrastlanması(MPK)) ve klinik-demografik özelliklere yönelik patolojik tam yanıt/patolojik
tam olmayan yanıt (pCR-non pCR grup) ile , uyumlu/uyumsuz grupta kıyaslamalı analiz yapıldı.
Ayrıca bu çalışmada karşı sağlıklı memede manuel olarak sınırlandırılan normal parankimde
MPK’nın kantitatif yarı otomatik analizi (bazal MRG ve kemoterapi sonrası) ile primer kitlenin
kantitatif yarı otomatik hacim ölçümü ve kitle içi sinyal yoğunluk (SI) değerleri kaydedildi. Tanı
anındaki MPK değerleri ile tümör yanıtı, menopoz durumu, yaş, yaş grubu, tümör fenotipi ve
tümör histolojik evresi arasındaki ilişki araştırıldı.
Bulgular: NAKT tamamlandıktan sonra 45 ( 48.4 %) lezyonda tam yanıt (pCR), 48 (51.6 %)
lezyonda tam olmayan yanıt (non pCR) saptanmış olup tüm çalışma grubunda ( p<0.001) , triple
negatif (p<0.001), luminal A (p=0.003), luminal B (p=0.042) ve HER2 reseptör pozitif
(p=0.048) alt gruplarında rCR-pCR arasında anlamlı istatistiki korelasyon bulunmuştur. Tüm
çalışma grubunda sensitivite % 82.2, spesifite % 66.7, doğruluk % 74.2, PPV ve NPV değerleri sırasıyla % 69.8 ve % 80’dir. Moleküler alt tiplere göre TN gruptan (%100) sonra en yüksek
sensitivite değerleri Luminal A (81.8%), Luminal B (88.7%) tümör grubu ile ER+ (84%)
olgularda izlenmiştir. Ayrıca NPV değerleri %90.5 ile Luminal A tip ve %90.0 ile PR(+) grupta
oldukça yüksek saptanmıştır.
Manuel ölçülen en uzun tümör çapı gruplar arasında anlamlı farklılık göstermez iken kantitatif
olarak ölçülen tedavi öncesi kitle hacim değerleri patolojik tam yanıt gösteren grupta daha
yüksek izlenmiştir (p: 0.021). Çalışma grubunda histolojik grade (MBR) ve/veya nükleer grade
skoru 3 olan lezyonların, daha düşük olan ( 1+2) lezyonlara göre daha düşük oranda tam yanıt
gösterdiği saptanmıştır ( p:0.005; p:0.018). Hormon reseptör pozitifliği mevcut grupta
NAKT’ye tam yanıt oranları hormon reseptör negatif gruba kıyasla daha düşük bulunmuştur ( p:
0.014). Tüm çalışma grubunda radyopatolojik uyumsuzluk oranı 24/93 (%25.8) olup uyumlu ve
uyumsuz grup arasında HER2 amplifikasyonu ve moleküler alt tiplere göre anlamlı istatistiki
farklılık olduğu izlenmiştir (p=0.048, p=0.026). Yanlış negatif 8 olgudan 4’ü HER2’den zengin
tipti. Tedavi öncesi ve sonrası MPK değerleri gruplar arası farklılık göstermezken pCR ve ER(+)
gruplarında tedavi sonrasında belirgin azalmıştır (p=0.001; p=0.027). Ayrıca tedavi sonrası
MPK değişim oranları diğer gruplarda (pCR/nonpCR; yas grubu ve menopoz durumu) farklılık
göstermezken ER(+) olgularda farklı bulunmuştur (p<0.001).
Sonuç:Tüm çalışma grubunda NAKT sonrası pCR tahmininde MRG yüksek doğruluk
gösterdi. pCR'yi tahmin etmede MRG'nin duyarlılık, özgüllük, NPV, PPV ve doğruluğu, meme
kanseri moleküler alt tipleri arasında önemli ölçüde farklılık göstermiş olup en yüksek NPV ve
sensitivite değerleri TN, Luminal A ve hormon reseptörü (HR) (+) grupta bulundu. Bu çalışma
HER2’den zengin tip ve tek başına HER2 reseptör pozitifliği durumunda MRG ile NAKT yanıtı
öngörü değerlerinin önemli ölçüde azaldığını ortaya koymuş olup bu olgularda MRG’nin pCR ve
non-pCR'yi tahmin etmek için tek başına yeterli bir yöntem olmadığını göstermiştir. Kesin
sonuçlara varmadan önce daha büyük olgu grupları gerekli olsa da , verilerimiz pCR için MRG ile
tahminin genel doğruluğunun artırılabileceğini önermektedir. Karşı sağlıklı memede tedaviye bağlı
MPK’da azalmayı kantitatif bir şekilde göstermiş olup , pCR-non pCR gruplarında ayrım için yarı
otomatik yöntemle yapılan bu çalışmada yeterli ayrım sağlayan bir belirteç olmadığını göstermiştir.
Bu yönden , otomatık-yarı otomatik ölçüm tekniklerinin, daha geniş hasta gruplarında
karşılaştırılması yarar sağlayacaktır.
Objectives: The aim of this study was to evaluate the accuracy of dynamic breast magnetic resonance imaging (MRI), in assessing the pathological response to neoadjuvant systemic chemotherapy (NACT) and determining histopathological, molecular and morphological features that affect the accuracy of response prediction. Methods: Our institutional review board approved this retrospective study. We included 90 patients (which 3 of them have bilateral breast cancer; mean age 46.03 ± 9.58; age range 24 years), with 93 invazive breast malign lesions, who underwent dynamic breast MR imaging before and after NACT. rCR was diagnosed when the original lesion site showed no enhancement in early dynamic phases. pCR was defined as the complete absence of invasive and microinvasive cancer in the breast parenchyma regardless of presence of DCIS. The axillary involvement wasn’t included in the radiopathological assessment. The diagnostic performance of dynamic MRI (accuracy, sensitivity, specificity, NPV, PPV), the factors affecting it (mass, size, volume, type, breast parenchyma enhancement) and clinical-demographic features were analyzed. A comparative analysis was performed in “pCR” and “non pCR” , “discordance” and “concordance” groups in predicting the response to NACT. Additionally, this study investigated the association between semiautomatic analysis of contralateral normal BPE that limited by ROI and semi-quantitative measurement of the primary mass volume and intramass SI values (before and after systemic chemotheraphy). Relationships between BPE and tumor response, menopausal status, age, age group, tumor phenotype, NAC type and tumor histological stage at the time of diagnosis were also evaluated. Results: After NACT, complete response (pCR) in 45 (48.4%) lesions and incomplete response (non pCR) in 48 (51.6%) lesions were detected. Ther was a significant statistical correlation in whole study group (p <0.001), triple negative (p <0.001), luminal A ( p = 0.003), luminal B (p = 0.042), and HER2 receptor positive (p = 0.048) subgroups, between rCR-pCR. In the whole study group, sensitivity was 82.2%, specificity 66.7%, accuracy 74.2%, PPV and NPV values were 69.8% and 80%, respectively. According to molecular subtypes, the highest sensitivity values after TN group (100%) were observed in Luminal A (81.8%), Luminal B (88.7%) tumor group and ER + (84%) cases. Furthermore, NPV values were found to quitely higher in Luminal A type with 90.5% and PR (+) group with 90.0%. Additionally, while the longest tumor diameter measured manually did not differ significantly between the groups, quantitatively measured pre-treatment mass volume was higher in pathological complete response group(p: 0.021). The lesions which have histological (MBR) and/or nuclear grade score 3 showed a lower complete response rate than the lesions lower grade (p: 0.005; p: 0.018). Hormone receptor positive lesions showed lower complete response rates to NACT in comparison HR(+) lesions (p: 0.014). The rate of radiopathological discordance in the whole study group was 24/93 (25.8%) and there was a significantly statistical difference between the concordant and discordant groups according to HER2 amplification and molecular subtypes (p = 0.048, p = 0.026). Four of eight false negative cases were HER2 rich type. MPK values decreased significantly in the pCR and ER (+) groups after treatment (p = 0.001; p = 0.027). But before and after treatment they did not differ between groups. In addition, while MPK changes after treatment did not differ in other groups (pCR / nonpCR; age group and menopause status) except ER (+) cases (p <0.001). Conclusion: MRI showed high accuracy to predict pCR after NACT, in all study groups. The sensitivity, specificity, NPV, PPV and accuracy of MRI in predicting pCR showed a significance differency among breast cancer molecular subtypes. Highest NPV and sensitivity rates were found in TN, Luminal A and hormone receptor (HR) (+) groups. This study showed that in the HER2-rich molecular type and HER2 receptor positivity cases, the predictive values of NACT response with MRI decreased significantly, and it showed that MRI alone is not a sufficient method to predict pCR and non-pCR in these cases. Our data suggest that the overall accuracy of the prediction with MRI for pCR can be increased, although larger case groups are required before reaching definitive conclusions. Furthermore, this study quantitatively showed the reduction in BPE due to treatment, in the contralateral healthy breast, and showed that there was not a marker that provides sufficient discrimination in this semi-automatic study for differentiation in pCR-non pCR groups. In this respect, it would be beneficial to compare automatic-semi-automatic measurement techniques in larger patient groups.
Objectives: The aim of this study was to evaluate the accuracy of dynamic breast magnetic resonance imaging (MRI), in assessing the pathological response to neoadjuvant systemic chemotherapy (NACT) and determining histopathological, molecular and morphological features that affect the accuracy of response prediction. Methods: Our institutional review board approved this retrospective study. We included 90 patients (which 3 of them have bilateral breast cancer; mean age 46.03 ± 9.58; age range 24 years), with 93 invazive breast malign lesions, who underwent dynamic breast MR imaging before and after NACT. rCR was diagnosed when the original lesion site showed no enhancement in early dynamic phases. pCR was defined as the complete absence of invasive and microinvasive cancer in the breast parenchyma regardless of presence of DCIS. The axillary involvement wasn’t included in the radiopathological assessment. The diagnostic performance of dynamic MRI (accuracy, sensitivity, specificity, NPV, PPV), the factors affecting it (mass, size, volume, type, breast parenchyma enhancement) and clinical-demographic features were analyzed. A comparative analysis was performed in “pCR” and “non pCR” , “discordance” and “concordance” groups in predicting the response to NACT. Additionally, this study investigated the association between semiautomatic analysis of contralateral normal BPE that limited by ROI and semi-quantitative measurement of the primary mass volume and intramass SI values (before and after systemic chemotheraphy). Relationships between BPE and tumor response, menopausal status, age, age group, tumor phenotype, NAC type and tumor histological stage at the time of diagnosis were also evaluated. Results: After NACT, complete response (pCR) in 45 (48.4%) lesions and incomplete response (non pCR) in 48 (51.6%) lesions were detected. Ther was a significant statistical correlation in whole study group (p <0.001), triple negative (p <0.001), luminal A ( p = 0.003), luminal B (p = 0.042), and HER2 receptor positive (p = 0.048) subgroups, between rCR-pCR. In the whole study group, sensitivity was 82.2%, specificity 66.7%, accuracy 74.2%, PPV and NPV values were 69.8% and 80%, respectively. According to molecular subtypes, the highest sensitivity values after TN group (100%) were observed in Luminal A (81.8%), Luminal B (88.7%) tumor group and ER + (84%) cases. Furthermore, NPV values were found to quitely higher in Luminal A type with 90.5% and PR (+) group with 90.0%. Additionally, while the longest tumor diameter measured manually did not differ significantly between the groups, quantitatively measured pre-treatment mass volume was higher in pathological complete response group(p: 0.021). The lesions which have histological (MBR) and/or nuclear grade score 3 showed a lower complete response rate than the lesions lower grade (p: 0.005; p: 0.018). Hormone receptor positive lesions showed lower complete response rates to NACT in comparison HR(+) lesions (p: 0.014). The rate of radiopathological discordance in the whole study group was 24/93 (25.8%) and there was a significantly statistical difference between the concordant and discordant groups according to HER2 amplification and molecular subtypes (p = 0.048, p = 0.026). Four of eight false negative cases were HER2 rich type. MPK values decreased significantly in the pCR and ER (+) groups after treatment (p = 0.001; p = 0.027). But before and after treatment they did not differ between groups. In addition, while MPK changes after treatment did not differ in other groups (pCR / nonpCR; age group and menopause status) except ER (+) cases (p <0.001). Conclusion: MRI showed high accuracy to predict pCR after NACT, in all study groups. The sensitivity, specificity, NPV, PPV and accuracy of MRI in predicting pCR showed a significance differency among breast cancer molecular subtypes. Highest NPV and sensitivity rates were found in TN, Luminal A and hormone receptor (HR) (+) groups. This study showed that in the HER2-rich molecular type and HER2 receptor positivity cases, the predictive values of NACT response with MRI decreased significantly, and it showed that MRI alone is not a sufficient method to predict pCR and non-pCR in these cases. Our data suggest that the overall accuracy of the prediction with MRI for pCR can be increased, although larger case groups are required before reaching definitive conclusions. Furthermore, this study quantitatively showed the reduction in BPE due to treatment, in the contralateral healthy breast, and showed that there was not a marker that provides sufficient discrimination in this semi-automatic study for differentiation in pCR-non pCR groups. In this respect, it would be beneficial to compare automatic-semi-automatic measurement techniques in larger patient groups.
Açıklama
Anahtar Kelimeler
NAK, Meme MR, Lokal İleri Meme Kanseri
