Table 1.

Summary of studies reporting on the role of IL6ST as a biomarker in breast cancer, including study cohorts and main findings. Studies are listed in chronological order of the original publication. See Table 2 for further description of the multifactor signatures. All the described associations achieved statistical significance (at least p < 0.05).

Original Publication	Study Type	Study Cohorts	Associations Reported	Main Predictive or Prognostic Value
Karczewska et al. (2000) [38]	Independent biomarker	75 PBCs who received surgery +/− adjuvant therapy.	IL6ST expression strongly correlates with earlier disease stages. In advanced stages, IL6ST expression is associated with better prognosis and higher OS and DFS rates. IL6ST negatively correlates with lymph node status and tumour size. IL6ST is independent from other well established clinicopathological factors.	IL6ST is a positive prognostic factor.
Tozlu et al. (2006) [43]	Independent biomarker	PBCs who received surgery (+ ET for ER+): -12 in screening set. -36 in validation set.	IL6ST is a perfect discriminator of ER+ status.	IL6ST is predictive for ER status and likely endocrine responsiveness.
Filipits et al. (2011) [44]	Molecular signatures: EP and EPclin	Original cohorts of ER+/HER2- BCs treated with ET: -964 in training set. -2948 in validation sets [44,45,46,47]. -ER+/HER2- BCs chemotherapy study [48]: -2630 in ET alone arm. -1116 in ET + chemotherapy arm.	EP and EPclin scores (linked to lower IL6ST expression) are continuous predictors of the risk of distant recurrence. EPclin is also prognostic for disease recurrence in patients who received chemotherapy, regardless of menopausal status. Patients with higher EPclin score derive benefit from the addition of chemotherapy to ET.	EP and EPclin stratify into risk groups that are prognostic for risk of distant recurrence at 5, 10 and 15 years in ER+/HER2- patients. EPclin is also prognostic for LRFS. EPclin high-risk group is predictive for chemotherapy benefit in pre- and postmenopausal ER+/HER2- patients.
Sota et al. (2014) [49]	Molecular signature: IRSN-23	PBCs who received NAC: -58 in training set. -59 in validation set. -901 in external validation set (publicly-available data).	Higher IL6ST is associated with lack of pCR from NAC. IRSN-23 classifies into Gp-R and Gp-NR groups, with differential response to NAC.	IRSN-23 signature stratifies into groups predictive of response to NAC, regardless of BC subtype of chemotherapy regimen.
Andres et al. (2014) [50]	Independent biomarker	Tumour marker analysis: -98 male BCs (publicly-available data). -18,366 female BCs (publicly-available data). -Gene expression analysis validation: -12 male BCs. -233 female BCs.	IL6ST expression is significantly elevated in male BCs compared to female malignancies. IL6ST correlates with ER expression.
Mathe et al. (2015) [40]	Independent biomarker	Screening set: -33 TNBCs; 17/33 with matched normal tissue, 15/33 with lymph node metastases. -Validation sets: -16 TNBCs; 4/16 with matched normal tissue -255 non-TNBC. -Independent validation sets [41]: -255 (publicly-available data) TNBCs. -148 TNBCs.	IL6ST expression is associated with longer survival. IL6ST expression is lower in TNBC than ER+ tumours.	IL6ST is prognostic for OS and RFS in TNBC.
Fertig et al. (2015) [42]	Independent biomarker	638 + 897 PBCs from publicly-available sets.	IL6ST expression is higher in luminal tumours (ER+/PR+) than in other BC subtypes. Positive trend towards longer survival in IL6ST+ luminal A tumours.
Turnbull et al. (2015) [51]	Molecular signatures: EER4, EA2 and EA2clin	EER4 cohort of ER+ postmenopausal IBCs treated with NET & ET: -73 training set. -44 validation set. -EA/EA2clin study cohort of ER+ IBCs treated with NET & ET [52,53]: -186 postmenopausal. -51 premenopausal.	IL6ST alone is an independent predictor of response to AIs. EER4 predicts response to AIs with greater accuracy and also predict RFS and BCSS. EA2 and EA2clin predict outcome from adjuvant ET with greater accuracy and also predict RFS and BCSS. EA2 also predicts outcome in premenopausal women. EA2clin predicts treatment response regardless of ET regimen.	IL6ST is an independent predictive marker for AI response in ER+/HER2- patients. EER4 further improves on this predictive ability. Models are prognostic of outcome (RFS, BCSS) from adjuvant ET response, regardless of menopausal status or ET regimen in ER+/HER2- patients.
Klahan et al. (2017) [39]	Independent biomarker	108 pretreated IBCs: -79 LVI+ -29 LVI-	IL6ST correlates with LVI in samples without lymph node metastasis and perineural invasion.
Tsunashima et al. (2018) [54]	Molecular signature: 42GC	ER+ BCs treated with ET who recurred: -177 training set (from publicly-available sets); 84 LR, 93 NLR. -201 validation set; 137 LR, 84 NLR.	Higher IL6ST is associated with lower risk of early recurrence but higher risk of late recurrence. 42GC classified intro LR and NLR groups, with differential risk of recurrence over time. could predict late recurrence	42GC stratifies into prognostic groups for risk of early and late recurrence in ER+ BC intervals.

42GC, 42-gene classifier; AI, aromatase inhibitor; BC, breast cancer; BCSS, BC-specific survival; DFS, disease-free survival; EA2, EndoAdjuvant 2; EA2clin, EndoAdjuvant 2 clinical; EER4, Edinburgh EndoResponse 4; EP, EndoPredict; EPclin, EndoPredict clinical; ER, oestrogen receptor; ET, endocrine therapy; Gp-NR, genomically-predicted non-responders; Gp-R, genomically-predicted responders; HER2, human epidermal growth factor receptor 2; IBC: invasive BC; IRSN-23, immune-related 23-gene signature for NAC; LN+: lymph node positive status; LR, late recurrence-like; LRFR, local recurrence-free survival; LVI, lympho-vascular invasion; NAC, neoadjuvant chemotherapy; NET, neoadjuvant ET; NLR, non-late recurrence-like; OS, overall survival; PBC, primary BC; pCR, pathological complete response; PR, progesterone receptor; RFS, recurrence-free survival; TNBC, triple negative BC.