Introduction: As tumour response rates are increasingly demonstrated in early-phase cancer trials (EPCT), optimal patient selection and accurate prognostication are paramount. Hammersmith Score (HS), a simple prognostic index derived on routine biochemical measures (albumin <35 g/L, lactate dehydrogenase >450 IU/L, sodium <135 mmol/L), is a validated predictor of response and survival in EPCT participants. HS has not been validated in the cancer immunotherapy era. Methods: We retrospectively analysed characteristics and outcomes of unselected referrals to our early-phase unit (12/2019–12/2022). Independent predictors for overall survival (OS) were identified from univariable and multivariable models. HS was calculated for 66 eligible trial participants and compared with the Royal Marsden Score (RMS) to predict OS. Multivariable logistic regression and C-index was used to compare predictive ability of prognostic models. Results: Of 212 referrals, 147 patients were screened and 82 patients treated in EPCT. Prognostic stratification by HS identifies significant difference in median OS, and HS was confirmed as a multivariable predictor for OS (HR: HS 1 vs. 0 2.51, 95% CI: 1.01–6.24, p = 0.049; HS 2/3 vs. 0: 10.32, 95% CI: 2.15–49.62, p = 0.004; C-index 0.771) with superior multivariable predictive ability than RMS (HR: RMS 2 vs. 0/1 5.46, 95% CI: 1.12–26.57, p = 0.036; RMS 3 vs. 0/1 6.83, 95% CI: 1.15–40.53, p < 0.001; C-index 0.743). Conclusions: HS is a validated prognostic index for patients with advanced cancer treated in the context of modern EPCTs, independent of tumour burden. HS is a simple, inexpensive prognostic tool to optimise referral for EPCT.

Significant anti-tumour responses in early-phase cancer trials (EPCTs) emphasize the need for precise patient selection and prognosis assessment. Prognostic scoring systems including the Royal Marsden Score (RMS) (albumin <35 g/L, lactate dehydrogenase [LDH] > ULN, >2 metastatic sites) [1] and Gustave-Roussy Immune Score (GRImS) (albumin <35 g/L, LDH > ULN, neutrophil-to-lymphocyte ratio >6) [2] quantify survival benefit, assigning a point for each parameter, with higher scores indicating shorter survival. The original Hammersmith Score (HS) multivariable analysis selected albumin <35 g/L, LDH >450 IU/L, and sodium (Na) <135 mmol/L as independent predictors for EPCT overall survival (OS) and offers a simple, cost-effective prognostic index, predicting response and survival in EPCTs [3].

Albumin levels reflect nutritional status, while lower levels may suggest malignancy-related inflammation, a poor prognostic factor [4]. LDH levels reflect tumour activity and anaerobic respiration, which are tied to rapid tumour growth [5]. Na has been observed as a prognostic indicator in early-phase trials, with this also seen in multiple tumour types [6].

HS and RMS were formulated prior to broad immunotherapy use in solid tumour oncology, with ipilimumab’s approval in 2011 heralding a new era in systemic anti-cancer treatment [7]. Unlike cytotoxics, immune checkpoint inhibitors have reproducible pharmacokinetics and non-linear dose-response relationships [8], and with the notable exception of ipilimumab, dose-limiting toxicity was not characterised for any of the forerunner ICIs [9]. This necessitates precise prognostication for uniform patient enrolment in EPCTs, given the varying responses at different exposure levels [10]. This study aimed to validate HS’s role as a prognostic marker for OS in EPCT participants, recognizing the evolving landscape of cancer therapies.

We conducted a retrospective analysis of all patients referred to the Hammersmith early-phase trials unit with solid malignancies, between December 2019 and December 2022. Patients were identified through electronic patient records, screening logs, and patient lists. Demographic and treatment data included gender, age, tumour type and number of metastatic sites, number of prior treatment lines, and biochemical and haematological test results. Radiological response data were collected, including imaging dates and treatment response, as per Response Evaluation Criteria in Solid Tumours (RECIST 1.1) criteria. The number of metastatic sites was quantified by a consultant radiologist.

Outcome measures (OS, cancer-specific) and PFS were calculated from the time of first treatment within an early-phase trial. Only patients who entered a trial within our unit were assigned a calculated RMS and HS and in turn were included in the analysis.

Statistical Analysis

Independent predictors for OS were assessed from univariable (Kaplan-Meier) and multivariable (Cox regression) analysis. HS was calculated for each patient and compared with the RMS for predicting OS.

Univariable cox regression analysis of component variables for each score was assessed for independent prognostic value within this new dataset. Multivariable analysis was assessed for independent prognostic ability. Discriminative ability of the HS and RMS for OS was calculated via C-index and compared with prognostic models.

Of 212 referrals, 147 patients were screened for early-phase clinical trials (EPCT), and 82 patients were treated and eligible to be assigned a HS and RMS. There were 65 screen failures and 2 patient refusals, and 63 patients did not enter a trial for unspecified reasons. At time of analysis, 59 patients had been withdrawn from an EPCT, 20 proceeded to receive standard-of-care treatment, 11 were re-treated in trials, and 24 were ineligible for post-progression treatment. Twenty-three patients remained on trial treatment at the time of analysis (Fig. 1a). Of 82 treated patients, 66 patients were assigned a HS and 66 patients were assigned an RMS – LDH was not measured in 12 patients, and albumin was not measured in 1 patient. In our cohort of patients (Table 1), median age was 62.1 years (range: 52.2–71.9), 45.1% patients were male, and performance status was ≥1 in a majority of patients (63.4%). The median number of metastatic sites was 2. Primary tumour groups included gynaecological (32), gastrointestinal (32), genitourinary (3), lung and mesothelioma (4), skin and melanoma (2), head and neck (6), and other types (3).

Fig. 1.

Prognostic factors in early-phase trial participants. a Sankey diagram illustrating patient referrals and outcomes. b Kaplan-Meier curve illustrating the comparison of OS according to HS. c Kaplan-Meier curve illustrating the comparison of OS as according to the RMS. PD, progressive disease; SoC, standard-of-care treatment.

Fig. 1.

Prognostic factors in early-phase trial participants. a Sankey diagram illustrating patient referrals and outcomes. b Kaplan-Meier curve illustrating the comparison of OS according to HS. c Kaplan-Meier curve illustrating the comparison of OS as according to the RMS. PD, progressive disease; SoC, standard-of-care treatment.

Close modal
Table 1.

Univariable and multivariable Cox regression analyses of prognostic factors for overall survival (OS)

Patient characteristicsUnivariableMultivariable
HR (95% CI)p valueaHR (95% CI)p value
Na <135 2.01 (0.76–5.29) 0.156  
Age ≥65 years 0.85 (0.42–1.72) 0.652  
ECOG PS ≥1 1.47 (0.717–3.05) 0.290  
Alb <35 3.46 (1.75–6.86) <0.001  
LDH >450 9.62 (2.74–33.87) <0.001  
No of metastatic sites ≥3 5.25 (2.33–11.80) <0.001 4.61 (1.55–13.71) 0.006 
Line of systemic treatment ≥3 2.95 (1.48–5.89) 0.002 0.61 (0.26–1.42) 0.255 
HS 
 0   
 1 3.52 (1.51–8.22) 0.004 2.51 (1.01–6.24) 0.049 
 2–3 20.09 (5.14–78.51) <0.001 10.32 (2.15–49.62) 0.004 
No of metastatic sites ≥3 5.25 (2.33–11.80) <0.001 0.50 (0.12–2.08) 0.342 
Line of systemic treatment ≥3 2.95 (1.48–5.89) 0.002 0.90 (0.39–2.10) 0.810 
RMS 
 0–1   
 2 8.38 (2.28–30.88) 0.001 5.46 (1.12–26.57) 0.036 
 3 13.15 (3.69–49.53) <0.001 6.83 (1.15–40.53) 0.035 
Patient characteristicsUnivariableMultivariable
HR (95% CI)p valueaHR (95% CI)p value
Na <135 2.01 (0.76–5.29) 0.156  
Age ≥65 years 0.85 (0.42–1.72) 0.652  
ECOG PS ≥1 1.47 (0.717–3.05) 0.290  
Alb <35 3.46 (1.75–6.86) <0.001  
LDH >450 9.62 (2.74–33.87) <0.001  
No of metastatic sites ≥3 5.25 (2.33–11.80) <0.001 4.61 (1.55–13.71) 0.006 
Line of systemic treatment ≥3 2.95 (1.48–5.89) 0.002 0.61 (0.26–1.42) 0.255 
HS 
 0   
 1 3.52 (1.51–8.22) 0.004 2.51 (1.01–6.24) 0.049 
 2–3 20.09 (5.14–78.51) <0.001 10.32 (2.15–49.62) 0.004 
No of metastatic sites ≥3 5.25 (2.33–11.80) <0.001 0.50 (0.12–2.08) 0.342 
Line of systemic treatment ≥3 2.95 (1.48–5.89) 0.002 0.90 (0.39–2.10) 0.810 
RMS 
 0–1   
 2 8.38 (2.28–30.88) 0.001 5.46 (1.12–26.57) 0.036 
 3 13.15 (3.69–49.53) <0.001 6.83 (1.15–40.53) 0.035 

Of 82 patients who received trial treatment, 10 (12%) were selected for EPCT on the basis of a clinically targetable mutation and/or specific biomarker of presumed susceptibility. Experimental therapies administered comprised immunotherapy (n = 53), targeted therapy (n = 6), and chemotherapy (n = 23).

LDH at cut-off values established for RMS (>243 IU/L, HR 9.62, 95% CI: 2.74–33.87, p < 0.001) and for HS (>450 IU/L, HR 4.18, 95% CI: 1.81–9.62, p < 0.001), number of metastatic sites ≥3 (HR 5.25 (2.33–11.80), p < 0.001), and line of systemic treatment (HR 2.95, 95% CI: 1.48–5.89, p < 0.002) independently predicted for OS. Na < 135 was not an independent predictor for OS (HR 2.01, 95% CI: 0.76–5.29, p = 0.156). In multivariable analysis alongside the number of metastatic sites ≥3 (HR 4.61 (1.55–13.71), p = 0.006) and line of systemic treatment (HR 0.61, 95% CI: 0.26–1.42, p = 0.255), both HS and RMS independently predicted for OS (Table 1).

Prognostic stratification based on HS (Fig. 1b) identified significantly different median OS: 11.28 months for HS 0 (95% CI: 8.34–14.21), 6.44 months (95% CI: 4.10–8.80) for HS 1, and 1.18 months for HS 2–3 (95% CI: 1.03–1.34, p < 0.001). Prognostic stratification based on RMS also demonstrated significantly different median OS: RMS 0–1: OS: 14.07 months (95% CI: 8.97–19.18), RMS 1: OS: 5.29 months (95% CI: 2.59–7.99), RMS 2–3: OS: 6.28 months (95% CI: 3.30–9.26), p < 0.001) (Fig. 1c). HS was confirmed as a multivariable predictor for OS (HR: HS 1 vs. 0 2.51, 95% CI: 1.01–6.24, p = 0.049; HS 2/3 vs. 0: 10.32, 95% CI: 2.15–49.62, p = 0.004; C-index 0.771) with similar multivariable predictive ability as RMS (HR: RMS 2 vs. 0/1 5.46, 95% CI: 1.12–26.57, p = 0.036; RMS 3 vs. 0/1 6.83, 95% CI: 1.15–40.53, p < 0.001; C-index 0.743).

When restricting to trial patients who received an immunotherapeutic investigational medicinal product (IMP), HS demonstrates multivariable predictive ability for OS though only HS2/3 versus HS0 reached significance (HR: HS 1 vs. 0 4.59, 95% CI: 0.78–27.19, p = 0.093; HS 2/3 vs. 0: 11.90, 95% CI: 1.55–91.30, p = 0.017; C-index 0.798 [0.708–0.887]), while RMS retained predictive ability across all scores (HR: RMS 2 vs. 0/1 8.37, 95% CI: 1.13–62.14, p = 0.038; RMS 3 vs. 0/1 30.47, 95% CI: 2.21–420.23, p = 0.011; C-index 0.798 [0.761–0835]) (Table 2).

Table 2.

Multivariable Cox regression analyses of prognostic factors for overall survival (OS) – immunotherapy (HS and RMS) and gastrointestinal and gynaecological subtypes (HS)

Patient characteristicsMultivariable
aHR (95% CI)p value
HS (immunotherapy treatment, n = 53) 
 0  
 1 4.59 (0.78–27.19) 0.093 
 2–3 11.90 (1.55–91.30) 0.017 
C-index 0.798 (0.708–0.887)  
RMS (immunotherapy treatment, n = 53) 
 0/1  
 2 8.37 (1.13–62.14) 0.038 
 3 30.47 (2.21–420.23) 0.011 
C-index 0.798 (0.761–0835)  
HS (gastrointestinal tumours, n = 32) 
 0  
 1 0.96 (0.24–3.78) 0.95 
 2–3 2.85 (0.26–31.81) 0.40 
HS (gynaecological tumours, n = 32) 
 0  
 1 1.00 (0.230–4.35) 1.00 
 2–3 1.00 (0.002–408.16) 1.00 
Patient characteristicsMultivariable
aHR (95% CI)p value
HS (immunotherapy treatment, n = 53) 
 0  
 1 4.59 (0.78–27.19) 0.093 
 2–3 11.90 (1.55–91.30) 0.017 
C-index 0.798 (0.708–0.887)  
RMS (immunotherapy treatment, n = 53) 
 0/1  
 2 8.37 (1.13–62.14) 0.038 
 3 30.47 (2.21–420.23) 0.011 
C-index 0.798 (0.761–0835)  
HS (gastrointestinal tumours, n = 32) 
 0  
 1 0.96 (0.24–3.78) 0.95 
 2–3 2.85 (0.26–31.81) 0.40 
HS (gynaecological tumours, n = 32) 
 0  
 1 1.00 (0.230–4.35) 1.00 
 2–3 1.00 (0.002–408.16) 1.00 

In the two largest tumour type cohorts, HS does not demonstrate significant multivariable predictive ability for OS in gastrointestinal-type (n = 32, 39%) (HR: HS 1 vs. 0 0.96, 95% CI: 0.24–3.78, p = 0.95; HS 2/3 vs. 0: 2.85, 95% CI: 0.26–31.81, p = 0.40) or gynaecological-type tumours (n = 32, 39%) (HR: HS 1 vs. 0 1.00, 95% CI: 0.230–4.35, p = 1.00; HS 2/3 vs. 0: 1.00, 95% CI: 0.002–408.16, p = 1.00) (Table 2).

Phase 1 trials require precise patient selection, with our analysis revealing a correlation between baseline characteristics and observed efficacy outcomes. HS was built upon a multivariable analysis conducted based on a trial portfolio comprising predominantly targeted agents and chemotherapy; therefore, the primary purpose of our study was to validate these prognostic indices in a contemporary, immunotherapy-majority cohort. Albumin, a surrogate marker for nutritional status and inflammation [11], independently predicts survival in EPCTs and features in the HS, RMS, and GRImS [12].

Elevated LDH, another independent predictor of survival in EPCTs, may signify immunosuppression in the tumour microenvironment [9], particularly relevant in a phase 1 cohort where 65% of patients received immunotherapy IMPs [9]. HS uses an LDH cut-off of 450 (HR 9.62), while RMH employs a >ULN (235, in our centre) cut-off (HR 4.18). This emphasis on LDH contributes to HS’s significance as an independent predictor of OS in a contemporary EPCT portfolio containing a comparatively larger proportion of immunotherapeutic IMPs, unlike the original validation cohort, which comprised targeted and cytotoxic therapies only [3].

In our univariable analysis, Na no longer independently predicted for OS in EPCTs. Although hyponatremia has correlated with OS in lung cancer [13] and solid tumours overall [14], the era of immunotherapy suggests a need for prognostic scores to emphasize biomarkers characterizing the tumour immune microenvironment. Indeed, substituting a lower >ULN LDH cut-off in HS removes its independence as an OS predictor in contemporary EPCTs (p = 0.929). LDH was inconsistently measured across protocols, which in turn affected the calculation of either prognostic score.

HS and RMS exhibit similar OS prediction abilities, though notably this did not carry across into an immunotherapeutic-only cohort. While >3 metastatic sites predict OS in EPCTs, with this phenomenon noted in a number of solid tumours [15‒17], introducing the need to quantify metastatic sites in RMS adds complexity compared to HS.

Notably, when separately assessing either gynaecological- or gastrointestinal-type tumours, HS was no longer a multivariable predictor for OS. However, our study has limitations, such as a small dataset (n = 82) and its retrospective, single-institution nature. Accurate assessment of tumour-specific prognostic effect is not possible in this small dataset, and therefore, only an assessment of prognosis across a general, heterogenous phase 1 cohort can be ascertained from this dataset, given the significant amount of censored cases.

Further studies outside of our institution would account for differing clinical practice, proportion of referable patient tumour groups, and availability of EPCT participants. However, extensive validation of HS and RMS in previous EPCTs lends credibility to our study’s generalisability.

In the immunotherapy era, HS remains a predictor of survival in EPCTs. The relevance however of prognostic scores varies with modern anti-cancer therapies. The HS originally chose variables based on responses to a predominantly targeted and chemotherapy-based phase 1 portfolio and as such vary in predictive ability when assessed in an immunotherapy-weighted cohort. Therefore, this study emphasizes the need to tailor prognostic indices to the treatments administered. The findings suggest a need to focus on identifying tumour immune microenvironment biomarkers for EPCT survival prediction with immunotherapies – alternative prognostic indices may be necessary for non-immune mechanisms.

In this retrospective analysis of a heterogenous phase 1 cohort, HS is validated as a prognostic index for advanced cancer patients in modern EPCTs, providing up to 6-fold survival differences regardless of tumour burden, simplifying patient referral. Further research is necessary for treatment-specific prognostic indices.

This retrospective research involved patients who provided explicit written informed consent for their data to be stored within the hospital’s database and utilized for scientific research. Ethical approval was granted by the Imperial College Tissue Bank (Reference Number: R16008). All clinical investigations were performed in adherence to the ethical principles detailed in the Declaration of Helsinki.

J.A.K. received travel grant funds from Boehringer-Ingelheim. D.J.P. received lecture fees from ViiV Healthcare, Bayer Healthcare, EISAI, BMS, and Roche; travel expenses from BMS and Bayer Healthcare; consulting fees from Mina Therapeutics, DaVolterra, Mursla, IPSEN, Exact Sciences, Avamune, EISAI, Roche, and AstraZeneca; received research funding (to institution) from MSD, GSK, and BMS. A.C. received grants for consultancies/advisory boards from BMS, MSD, OncoC4, IQVIA, Roche, GSK, AstraZeneca, Access Infinity, Ardelis Health and received speaker’s fees from AstraZeneca, EISAI, Pierre-Fabre, MSD. B.S. received travel support from AbbVie, Ipsen, AstraZeneca, and Gilead and research funding (to institution) from Eisai. A.D. received educational support for congress attendance and consultancy fees from Roche. L.T. served on advisory boards for AstraZeneca and Clovis Oncology; received consulting fees from Tesaro, AstraZeneca, GSK, and Clovis Oncology; received honoraria for lectures and presentations from Clovis Oncology, GSK, AstraZeneca, and MSD; and sponsorship for travel and conference attendance from Tesaro and AstraZeneca.

D.J.P. is supported by grant funding from the Wellcome Trust Strategic Fund (PS3416), the Associazione Italiana per la Ricerca sul Cancro (AIRC MFAG 25697) and acknowledges grant support from the Cancer Treatment and Research Trust (CTRT) and infrastructural support by the Imperial Experimental Cancer Medicine Centre and the NIHR Imperial Biomedical Research Centre. A.C. is supported by the NIHR Imperial BRC. D.J.P. has received direct project funding by the NIHR Imperial Biomedical Research Centre (BRC), ITMAT Push for Impact Grant Scheme 2019. The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care. A.D. is supported by the National Institute for Health Research (NIHR) Imperial BRC, by grant funding from the European Association for the Study of the Liver (2021 Andrew Burroughs Fellowship) and from Cancer Research UK (RCCPDB- Nov21/100008).

J.A.K., B.S., and D.J.P. contributed to the writing and editing of this manuscript and conceived and designed the experiments. Data analysis was performed by J.A.K., B.S., A.D., A.C., and C.A.M.F. C.P., A.M., S.P., D.M.G., W.M., O.M., A.G., L.B., S.C., T.P., W.S., J.R., Y.N., J. Krell, I.M., L.T., W.-H.E.P., M.A., and J.S.E. reviewed and edited the final version of the manuscript.

The data that upon which the findings of this study are based are available upon reasonable written request to the corresponding author, J.A.K., and are in secure storage at Hammersmith Hospital. The data are not publicly available due to the presence of information that could compromise the privacy of research participants.

1.
Arkenau
HT
,
Barriuso
J
,
Olmos
D
,
Ang
JE
,
de Bono
J
,
Judson
I
, et al
.
Prospective validation of a prognostic score to improve patient selection for oncology phase I trials
.
J Clin Oncol
.
2009
;
27
(
16
):
2692
6
.
2.
Bigot
F
,
Castanon
E
,
Baldini
C
,
Hollebecque
A
,
Carmona
A
,
Postel-Vinay
S
, et al
.
Prospective validation of a prognostic score for patients in immunotherapy phase I trials: the Gustave Roussy Immune Score (GRIm-Score)
.
Eur J Cancer
.
2017
;
84
:
212
8
.
3.
Stavraka
C
,
Pinato
DJ
,
Turnbull
SJ
,
Flynn
MJ
,
Forster
MD
,
O'Cathail
SM
, et al
.
Developing an objective marker to optimize patient selection and predict survival benefit in early-phase cancer trials
.
Cancer
.
2014
;
120
(
2
):
262
70
.
4.
Stares
MD
,
Swan
A
,
Cumming
K
,
Ding
TE
,
Leach
J
,
Stratton
C
, et al
.
Hypoalbuminaemia as a prognostic biomarker of first-line treatment resistance in metastatic non-small cell lung cancer
.
Front Nutr
.
2021
;
8
:
734735
.
5.
Wulaningsih
W
,
Holmberg
L
,
Garmo
H
,
Malmstrom
H
,
Lambe
M
,
Hammar
N
, et al
.
Serum lactate dehydrogenase and survival following cancer diagnosis
.
Br J Cancer
.
2015
;
113
(
9
):
1389
96
.
6.
Roviello
G
,
Catalano
M
,
De Giorgi
U
,
Maruzzo
M
,
Buti
S
,
Gambale
E
, et al
.
Prognostic value of normal sodium levels in patients with metastatic renal cell carcinoma receiving tyrosine kinase inhibitors
.
Front Oncol
.
2022
;
12
:
918413
.
7.
Eno
J
.
Immunotherapy through the years
.
J Adv Pract Oncol
.
2017
;
8
(
7
):
747
53
.
8.
Le Louedec
F
,
Leenhardt
F
,
Marin
C
,
Chatelut
É
,
Evrard
A
,
Ciccolini
J
.
Cancer immunotherapy dosing: a pharmacokinetic/pharmacodynamic perspective
.
Vaccines
.
2020
;
8
(
4
):
632
. Published 2020 Oct 31.
9.
Viala
M
,
Vinches
M
,
Alexandre
M
,
Mollevi
C
,
Durigova
A
,
Hayaoui
N
, et al
.
Strategies for clinical development of monoclonal antibodies beyond first-in-human trials: tested doses and rationale for dose selection
.
Br J Cancer
.
2018
;
118
(
5
):
679
97
.
10.
Fourie Zirkelbach
J
,
Shah
M
,
Vallejo
J
,
Cheng
J
,
Ayyoub
A
,
Liu
J
, et al
.
Improving dose-optimization processes used in oncology drug development to minimize toxicity and maximize benefit to patients
.
J Clin Oncol
.
2022
;
40
(
30
):
3489
500
.
11.
Oñate-Ocaña
LF
,
Aiello-Crocifoglio
V
,
Gallardo-Rincón
D
,
Herrera-Goepfert
R
,
Brom-Valladares
R
,
Carrillo
JF
, et al
.
Serum albumin as a significant prognostic factor for patients with gastric carcinoma
.
Ann Surg Oncol
.
2007
;
14
(
2
):
381
9
.
12.
Qi
J
,
Gu
C
,
Wang
W
,
Xiang
MQ
,
Chen
X
,
Fu
J
.
Elevated lactate dehydrogenase levels display a poor prognostic factor for Non-Hodgkin’s lymphoma in intensive care unit: an analysis of the MIMIC-III database combined with external validation
.
Front Oncol
.
2021
;
11
:
753712
.
13.
Bartalis
E
,
Gergics
M
,
Tinusz
B
,
Földi
M
,
Kiss
S
,
Németh
D
, et al
.
Prevalence and prognostic significance of hyponatremia in patients with lung cancer: systematic review and meta-analysis
.
Front Med
.
2021
;
8
:
671951
.
14.
Fucà
G
,
Mariani
L
,
Lo Vullo
S
,
Galli
G
,
Berardi
R
,
Di Nicola
M
, et al
.
Weighing the prognostic role of hyponatremia in hospitalized patients with metastatic solid tumors: the HYPNOSIS study
.
Sci Rep
.
2019
;
9
(
1
):
12993
.
15.
Oh
YK
,
Taylor
S
,
Bekele
BN
,
Debnam
JM
,
Allen
PK
,
Suki
D
, et al
.
Number of metastatic sites is a strong predictor of survival in patients with nonsmall cell lung cancer with or without brain metastases
.
Cancer
.
2009
;
115
(
13
):
2930
8
.
16.
Bilen
MA
,
Shabto
JM
,
Martini
DJ
,
Liu
Y
,
Lewis
C
,
Collins
H
, et al
.
Sites of metastasis and association with clinical outcome in advanced stage cancer patients treated with immunotherapy
.
BMC Cancer
.
2019
;
19
(
1
):
857
.
17.
Zou
Y
,
Yang
QK
,
Wu
Y
,
Ai
H
,
Yao
Z
,
Zhang
C
, et al
.
Prognosticators and prognostic nomograms for leiomyosarcoma patients with metastasis
.
Front Oncol
.
2022
;
12
:
840962
.