1

Title: Quality of life and treatment response among women with platinum-resistant versus 1

platinum-sensitive ovarian cancer treated for progression: A prospective analysis. 2

3

Authors: Vanessa L Beesley1, Adele C Green2, David K Wyld3, Peter O’Rourke4, Leesa F 4

Wockner4, Anna deFazio5, Phyllis N Butow6, Melanie A Price6, Keith R Horwood7 5

Alexandra M Clavarino8, Australian Ovarian Cancer Study Group1, Australian Ovarian 6

Cancer Study–Quality of Life Study Investigators7, Penelope M Webb1 7

8

Affiliations: 9

1Gynaecological Cancers Group, QIMR Berghofer Medical Research Institute, Herston, 10

QLD, Australia 11

2Cancer and Population Studies Group, QIMR Berghofer Medical Research Institute, 12

Herston, QLD, Australia 13

3Department of Medical Oncology, Royal Brisbane and Women’s Hospital, Herston, QLD, 14

Australia 15

4Statistics Unit, QIMR Berghofer Medical Research Institute, Herston, QLD, Australia 16

5 Westmead Institute of Cancer Research, University of Sydney at Westmead Millennium 17

Institute and Westmead Hospital, Sydney, NSW, Australia 18

6Centre for Medical Psychology and Evidence-Based Decision-Making, School of 19

Psychology, The University of Sydney, Sydney, NSW, Australia 20

7Greenslopes Specialist Centre, Greenslopes Private Hospital, Greenslopes, QLD, Australia 21

8School of Pharmacy, University of Queensland, St. Lucia, QLD, Australia 22

23

24

2

Correspondence to: 25

Dr Vanessa L Beesley 26

QIMR Berghofer Medical Research Institute, 27

Gynaecological Cancers Group, 28

Locked Bag 2000, Royal Brisbane Hospital, 29

Herston, QLD 4029, Australia. 30

E-mail: Vanessa.Beesley@qimr.edu.au 31

Tel: +61 7 3362 0270 32

Fax: +61 7 3845 3503 33

34

35

36 37

38

3

Abstract: 39

Objective. Most women with ovarian cancer relapse and undergo further chemotherapy 40

however evidence regarding the benefits of this for women with platinum-resistant disease is 41

limited. Our objective was to determine whether there was a quality of life improvement or 42

treatment response among women treated for platinum-resistant recurrent ovarian cancer. 43

Methods. We combined data from 2 studies where women treated with chemotherapy for 44

recurrent ovarian cancer (n=172) completed a quality of life questionnaire every 3 months. 45

Cancers were classified as platinum-resistant if they progressed within 6 months of 46

completing first-line chemotherapy. Mixed effects models were used to analyse change in 47

quality of life during the first 6 months after second-line chemotherapy. 48

Results. One-quarter of women (n=44) were classified as having platinum-resistant disease. 49

Overall, their quality of life did not significantly increase or decrease, following 50

commencement of second-line chemotherapy (least square mean scores = 107, 105, 103 at 51

chemotherapy start, 3 and 6 months later, respectively), although 26% of these women 52

reported a meaningful increase and 31% reported a meaningful decline. One-third of the 53

platinum-resistant group responded (11% complete and 21% partial response) to second-line 54

chemotherapy, and this figure increased to 54% among the subset (36%) re-treated with 55

platinum-based agents with or without other agents. Preliminary analyses suggest quality of 56

life may be higher at chemotherapy initiation in women whose disease responded (median 57

score 121 vs 110). 58

Conclusions. Overall, quality of life appears to be maintained in women with platinum-59

resistant ovarian cancer who receive further chemotherapy and some women respond to re-60

treatment. 61

Keywords: ovarian cancer, oncology, progression, recurrence, chemotherapy, platinum-62

resistance, quality of life 63

4

64

Introduction 65

66

Ovarian cancer is the fifth most common cause of cancer death in women [1]. Two-thirds of 67

women diagnosed with ovarian cancer present with advanced disease and, although most 68

initially respond to treatment, the vast majority eventually experience disease progression [2]. 69

With few exceptions, recurrent ovarian cancer is not curable, yet end-points of most clinical 70

trials in this group include response to treatment, progression-free survival and overall 71

survival. In such a palliative setting, quality of life may be a more meaningful end-point. 72

73

After first-line chemotherapy for ovarian cancer, time to progression can be used to predict 74

how groups of patients might respond to further chemotherapy at relapse [3]. A cancer with a 75

long progression-free interval after completion of chemotherapy (>6 months) is associated 76

with higher response rates to re-treatment, and is classified as ‘platinum-sensitive’. 77

Management of recurrent disease in this group generally involves further platinum-based 78

chemotherapy with response rates ranging from 30-60% and a median progression-free 79

survival of 6-12 months [4, 5]. 80

81

Women whose disease progresses within 6 months of completing first-line chemotherapy are 82

defined as having ‘platinum-resistant’ disease. They have a 10-25% likelihood of responding 83

to second-line chemotherapy, a median progression-free survival of 3.5-4 months and a 84

median overall survival of 9-10 months [4, 6]. ‘Platinum-refractory’ disease refers to disease 85

that progresses during chemotherapy and the objective response rate to second-line 86

chemotherapy in these patients is <10% [4]. While a subset of women in these groups derive 87

a survival benefit from second-line chemotherapy, some women with platinum-88

5

resistant/refractory ovarian cancer continue to receive aggressive and expensive 89

chemotherapy until days before their death [7]. The goal of end of life care is to maximize 90

quality of life yet to date only 1 small study (n=9 with complete data) with a highly selected 91

patient group of women with good performance status, has considered quality of life 92

specifically among women with platinum-resistant recurrent ovarian cancer [8]. Other studies 93

which have combined women with platinum-sensitive and resistant disease, found quality of 94

life improvements after chemotherapy [9], however these pooled results are likely to be 95

driven by the high proportion of women with platinum-sensitive disease. We have used 2 96

population-based longitudinal datasets to determine if there is a quality of life improvement 97

following second-line chemotherapy among women with platinum-resistant ovarian cancer 98

and have compared this group to women with platinum-sensitive disease. We also consider 99

treatment response among these subgroups. 100

101

Methods 102

103

Study designs 104

105

We combined data from 2 longitudinal studies of women with ovarian cancer (Figure 1). 106

Ethics approval was obtained from the Human Research Ethics Committees of the QIMR 107

Berghofer Medical Research Institute, The University of Sydney and all participating 108

hospitals. Participants from both studies provided informed consent. 109

110

The first study, PROSPECT [10], recruited 74 women with newly-diagnosed and 48 with 111

recurrent disease from 7 hospitals in 2 Australian states between April 2003 and January 112

2005. Eligible patients were women aged 18 to 79 years referred for chemotherapy for 113

6

primary epithelial ovarian cancer and who were able to complete the study documents in 114

English. Participants were mailed questionnaires every 3 months for 2 years or until they 115

withdrew or died. At study entry participants were between 1 month and 20 years post-116

diagnosis (median 15 months). 117

118

The second study included 798 women who had participated in a previous population-based 119

case-control study [11-13]. The Australian Ovarian Cancer Study (AOCS) recruited women 120

aged 18–79 years with invasive epithelial ovarian cancer diagnosed between January 2002 121

and June 2006 through gynaecological oncology units and state-based cancer registries in all 122

Australian states and territories. Women who participated in AOCS and were still alive 123

between May 2005 and March 2007 were invited to take part in a quality of life substudy 124

(AOCS-QoL) [14]. Women who consented were mailed questionnaires at 3-monthly intervals 125

for up to 2 years, beginning 3 months to 5 years after diagnosis (median 26 months). 126

127

Women from the PROSPECT and AOCS-QoL studies were excluded from this analysis if 128

they did not experience a progression of their disease during data collection (n=434), or did 129

not complete a quality of life questionnaire within 2 months of starting chemotherapy for 130

their first disease progression (n=304, including 2 with platinum-resistant and 7 with 131

platinum-sensitive disease who did not have further chemotherapy) or if their disease was not 132

classifiable as platinum-resistant or –sensitive (n=7 of whom 2 were platinum-refractory). We 133

also removed duplicate data for 3 women who were in both studies leaving 172 women for 134

analysis. 135

136

7

Data collection instruments and classification 137

138

Platinum status and treatment response: We reviewed medical records for date of diagnosis, 139

start and completion dates of each chemotherapy course and cancer antigen (CA) -125 blood 140

levels at each follow-up with the treating doctor. Platinum status was classified based on 141

response to first-line treatment (CA-125 and CT scans when available) [4, 15, 16]. Women 142

who had at least a partial response to first-line platinum-based chemotherapy and no disease 143

progression within the first 6 months after completion of chemotherapy, were classified as 144

platinum-sensitive and those who had no response and/or disease progression within 6 145

months of completing chemotherapy were classified as platinum-resistant. Response to 146

chemotherapy after second-line treatment was determined based on CA-125 criteria alone. 147

That is, compared to a pretreatment CA-125, a complete response was defined as 148

normalization of CA-125 maintained for 28 days, a partial response was defined as at least a 149

50% reduction maintained for 28 days and no response was defined as no sustained reduction 150

in CA-125. 151

152

Quality of Life: This was measured using the 39-item Functional Assessment of Cancer 153

Therapy—Ovarian (FACT-O). The FACT-O is a multi-dimensional, ovarian cancer-specific 154

instrument, assessing 4 general sub-scales (physical, social, emotional and functional well-155

being) plus an ovarian cancer and treatment specific subscale. Overall quality of life is 156

derived from these 5 subscales. This instrument has undergone extensive reliability and 157

validity testing showing that internal consistency and test-retest reliability are adequate [17]. 158

159

Covariates: At baseline we collected self-reported demographic information including: date 160

of birth, marital status and education level. We also collected clinical data from medical 161

8

records including: International Federation of Gynecology and Obstetrics (FIGO) disease 162

stage at diagnosis and agents/regimen of each chemotherapy course. 163

164

Survival: Dates of disease progression and death were obtained from medical records. 165

Clinical follow-up of progression and survival status continued until September 2010 for 166

PROSPECT and July 2010 for AOCS. Survival time was calculated as the date of death, or if 167

alive, end of clinical follow-up period, minus the date of first progression. 168

169

Statistical methods 170

171

We analyzed data collected during the first 6 months after second-line chemotherapy. After 172

this there were fewer than 20 participants remaining per group. For the 14% of woman who 173

were missing data at the 3 or 6 month data-points yet remained active in the study, we 174

imputed the value as the average score of the time-points before and after the missing time-175

point. We documented reasons for dropout and calculated, at each time-point, the proportion 176

of women who dropped out due to death or incapacity (i.e. not missing completely at random) 177

by platinum status. To determine the likely effect of missingness [18] we graphed quality of 178

life scores by patterns of dropout, stratified by platinum status. To determine predictors of 179

missingness [18] we cross-tabulated the characteristics of those who dropped out versus those 180

who did not. 181

182

We used descriptive statistics, stratified by platinum status, to calculate the proportion of 183

women re-treated with different chemotherapy regimens and who had a response to 184

chemotherapy. Cox regression survival analysis was used to obtain hazard ratios and 95% 185

confidence intervals for post-progression survival by platinum status after adjustment for age. 186

9

187

A mixed effects repeated measures model with a random intercept and random slope was 188

used to estimate the least squares mean scores of quality of life and wellbeing subscales by 189

platinum status at baseline (0-2 months after the start of second-line chemotherapy) and 3 and 190

6 months later. We fitted the model using terms for platinum status, the interaction of 191

platinum status with time, and covariates including age at progression and whether a woman 192

was on chemotherapy at the time she completed the quality of life questionnaire. An 193

unstructured covariance matrix for the random effects was used as it had the best fit for this 194

dataset. Other key demographic and clinical covariates including marital status, education 195

level, second-line chemotherapy with a platinum-based drug (yes, no) and second-line 196

chemotherapy with liposomal doxorubicin (yes, no) were not included in the final models as 197

they were not significantly associated with quality of life in fully adjusted analyses. Although 198

not significant, age was retained as a standard covariate. Disease stage at diagnosis was not 199

included as almost all participants (96%) were classified as late stage. 200

201

We repeated our analyses excluding women (a) who dropped out due to death or incapacity 202

(n=31) and (b) with baseline data only (n=44), to assess the robustness of the results when 203

different data distributions were included. Furthermore, as all models have assumptions 204

about missing data we also conducted sensitivity analyses by fitting a repeated measures 205

ANOVA model with last observation carried forward and a weighted generalized estimating 206

equations model to see if the estimates were different from the mixed effects model. The 207

results across all models were similar and thus the estimates were deemed robust (data not 208

shown). 209

10

210

Clinically important changes and group differences (±) in outcomes were defined a priori as 211

follows: 8 for the overall FACT-O scale, 2 for the FACT-O wellbeing subscales and 2.5 for 212

cancer-specific concerns. These minimal differences were part way between the conservative 213

half standard deviation guideline [19] and what others have reported as clinically meaningful 214

differences in the gynecological cancer setting [20]. P-value were considered at the 215

conventional p<0.05 level. 216

217

Results 218

219

Participants 220

221

On average, women in our analysis were 60 years of age at diagnosis (SD = 10), most (71%) 222

were married or living with their partner, approximately three-quarters (74%) were classified 223

as platinum-sensitive and almost all had late stage disease at diagnosis (96%) (Table 1). 224

225

Characterizing missingness 226

227

Forty percent of women dropped out of the study due to death, incapacity, unknown reason or 228

because they had completed their commitment to the parent study (Figure 1). The proportion 229

of dropouts due to death or incapacity was lower in the platinum-sensitive group (33%) than 230

in the platinum-resistant group (75%). Irrespective of platinum status, women who dropped 231

out for any reason had lower mean quality of life at baseline than those with complete data 232

(Figure 2). Those who dropped out at the 6-month follow-up had similar absolute changes in 233

quality of life prior to that time-point to the women with complete data (Figure 2). 234

11

235

There were no systematic differences in the characteristics of the 37 women who dropped out 236

due to study completion or for unknown reasons and the 104 women with complete data 237

(Table 1). However the 31 women who dropped out due to death or incapacity were less 238

likely to be platinum-sensitive, less likely to have a partner, had lower baseline quality of life 239

and had shorter survival time compared to those with complete data (Table 1). 240

241

Second-line chemotherapy agents, responses and survival by platinum-status 242

243

Most women (87%, n=111) with platinum-sensitive disease were re-treated with a platinum-244

based regimen (Figure 3). Approximately one-third (36%, n=16) of women with platinum-245

resistant disease were re-treated with platinum with or without other chemotherapy agents, 246

while just under half (43%, n=19) were re-treated with liposomal doxorubicin (Figure 3). 247

248

Most women (84%, n=86/102 with adequate CA-125 data for evaluation) with platinum-249

sensitive disease had a CA-125 response to second line chemotherapy (complete response 250

65%, n=66; partial response 19%, n=20) compared to only 32% (n=12/38) of women with 251

platinum-resistant disease (complete response 11%, n=4; partial response 21%, n=8). Of the 252

women re-treated with platinum-based agents, 88% (n=79/90) of those who were platinum-253

sensitive and 54% (n=7/13) of those who were platinum-resistant, had a complete or partial 254

response. Many of these women also had other agents combined with platinum that they had 255

not received before. Five of the 7 women with platinum-resistant disease who had a response 256

were re-treated with platinum plus paclitaxel, docetaxel, gemcitabine or liposomal 257

doxorubicin. 258

259

12

Overall a total of 119 of the 172 women died during the clinical follow-up period; comprising 260

83 of the 128 women with platinum-sensitive disease and 36 of the 44 women with platinum-261

resistant disease. Women who had platinum-resistant disease had significantly worse post-262

progression survival than women who had platinum-sensitive disease (multivariate HR 2.34, 263

95% CI 1.57-3.49, p<0.001; median post-progression survival 13 versus 30 months). Among 264

women with platinum-resistant disease, median post-progression survival was 21 months for 265

the 16 women re-treated with platinum (and 31 months among the 7 women who responded 266

to re-treatment) versus 10 months for the 28 women not re-treated with platinum. 267

268

Quality of life changes during and after second-line chemotherapy 269

270

Among women with follow-up data and platinum-sensitive disease, 51% (n=49) had an 271

increase, 40% (n=39) had no meaningful change and 9% (n=9) experienced a decrease in 272

quality of life over the 6 months after starting second-line chemotherapy. In contrast, among 273

women with follow-up and platinum-resistant disease, 26% (n=8) had an increase, 42% 274

(n=13) had no meaningful change and 31% (n=10) reported a decrease in quality of life. 275

276

Group differences in quality of life at chemotherapy initiation 277

278

An unadjusted comparison among women with platinum-resistant disease and follow-up data 279

indicated that quality of life at chemotherapy initiation was no different in the 10 women who 280

experienced a decline in quality of life compared to the 21 who had preserved or improved 281

quality of life (median score 114 vs 112). However, among women with platinum-resistant 282

disease median quality of life at chemotherapy initiation was higher (median score 121 vs 283

110) in women whose disease responded to treatment (n=12) compared to women whose 284

13

disease progressed (n=26). 285

286

Platinum status as a predictor of quality of life during and after second-line chemotherapy 287

288

Mixed effects models revealed a clinically meaningful and statistically significant group 289

improvement in quality of life over the 6 months after starting second-line chemotherapy 290

among women with platinum-sensitive disease, whereas there was no significant or 291

precipitate decline among those with platinum-resistant disease (figure 4a). These results 292

were replicated in sensitivity analyses (figure 4b & 4c). The changes in quality of life over 293

time were statistically significantly different between the 2 groups in all models (p<0.05). 294

Current receipt of chemotherapy was the only other factor that was significantly associated 295

with (decreased) quality of life (p=0.008). Marital status, education level, platinum-based and 296

liposomal doxorubicin second-line chemotherapy agents were not significantly associated 297

with quality of life. 298

299

Within the quality of life subscales, a clinically significant but not statistically significant 300

reduction in physical wellbeing was reported over time in women with platinum-resistant 301

cancer, whereas a clinically and statistically significant increase in physical and functional 302

wellbeing and an improvement in ovarian cancer-specific symptoms was reported in women 303

with platinum-sensitive disease (Figure 5). 304

305

Discussion 306

307

This is the first population-based study to evaluate quality of life separately in women 308

classified as having platinum-sensitive and platinum-resistant disease. We have shown that, 309

14

overall, among women classified as having platinum-resistant disease, retreatment with 310

chemotherapy when they experience a recurrence does not significantly improve or diminish 311

their quality of life. However, one-quarter and one-third of this group respectively report 312

meaningful increases and decreases in quality of life. Like others [5, 6] we confirmed that for 313

women with platinum-resistant ovarian cancer, complete tumor response is rare and survival 314

is significantly poorer than for women with platinum-sensitive disease. Importantly though, a 315

subset of women with apparently platinum-resistant tumors do respond to re-treatment. 316

Notably, as reported here and elsewhere [21], the response in women with platinum-resistant 317

disease appears to be better for those re-treated with platinum-based agents ± other agents 318

than the response in those re-treated with liposomal doxorubicin or taxane alone. Larger 319

studies adjusting for characteristics of women selected to receive platinum are needed to 320

confirm this result and to determine if there is also a quality of life improvement or 321

maintenance. 322

323

In women with platinum-resistant disease maintenance of quality of life may be a useful 324

outcome, as there is a possibility that without chemotherapy quality of life would have 325

rapidly declined due to disease progression. This conclusion however is to be taken with 326

caution. We were unable to compare the quality of life trajectory of women not given second-327

line chemotherapy following disease progression since almost all women went on to receive 328

second-line chemotherapy. We also show that aside from platinum-status the only other 329

factor that was associated with improved quality of life within our overall sample was 330

cessation of chemotherapy. Furthermore, our analyses may underestimate decline in quality 331

of life among women with platinum-resistant disease as computation of missingness in the 332

mixed effects model is based on the observed data distribution and those women who 333

dropped out of the study due to incapacity were likely to have poorer quality of life. 334

15

335

Our results highlight the need for detailed investigation into which women with platinum-336

resistant ovarian cancer will and will not benefit from chemotherapy in the later stages of 337

their disease course. Our preliminary subgroup analysis gives an indication that quality of life 338

at the start of chemotherapy may be correlated with having a response to treatment. If this 339

were to be confirmed with a larger sample it is possible that quality of life screening at the 340

start of treatment may be a useful clinical practice in this particular group of patients. This is 341

important preliminary information since the decision to re-treat women with platinum-342

resistant disease may come at considerable cost, both in terms of increased patient morbidity, 343

and to the health system in terms of expenditure which is estimated at USD $20,744 on 344

average per woman after first-line chemotherapy [22]. 345

346

In conclusion, this study raises important ethical and economic questions about the re-347

treatment of women with platinum-resistant ovarian cancer with chemotherapy. Disease 348

response rates are particularly low in those not treated with platinum and approximately one 349

third of women in the platinum-resistant group when treated with second-line chemotherapy 350

have significant declines in quality of life. Further research in a larger series of patients with 351

platinum-resistant disease should be conducted to identify the groups most likely to benefit 352

from re-treatment with chemotherapy. Given that the majority of women with ovarian cancer 353

will have a progression and that approximately one-quarter will be classified as having 354

platinum-resistant disease, evaluation of both the costs and benefits of treating this disease is 355

essential for efficient use of limited health care resources and optimal patient-centered care 356

[23]. 357

358

16

Acknowledgments 359

Data for this analysis came from (1) the PROSPECT study which was funded by Cancer 360

Council Queensland, (2) The Australian Ovarian Cancer Study which was supported by the 361

US Army Medical Research and Materiel Command under DAMD17-01-1-0729, The Cancer 362

Councils of New South Wales, Queensland, South Australia, Tasmania and Victoria, The 363

Cancer Foundation of Western Australia and the National Health and Medical Research 364

Council (NHMRC) of Australia (199600 and 400413) and (3) The AOCS-QoL study which 365

was funded by the Cancer Councils of New South Wales and Queensland (RG 36/05). 366

367

Vanessa Beesley was funded by an NHMRC early career fellowship, Phyllis Butow and 368

Penelope Webb by NHMRC senior research fellowships, Melanie Price by a post-doctoral 369

fellowship from the School of Psychology, The University of Sydney and Alexandra 370

Clavarino by NHMRC Capacity Building Grant. 371

372

We thank Lynley Aldridge, Diana Grivas, Naomi McGowan, Ann Ward, Colleen Loos, 373

Rebekah Cicero and Alessandra Francesconi for their help with the data tracking, 374

classification and management for the PROSPECT and AOCS-QoL studies. We also 375

acknowledge all participating institutions and investigators represented within the AOCS 376

group listed at www.aocstudy.org. 377

378

Conflicts of interests 379

A. deFazio holds a patent unrelated to this manuscript and has received a Speaker's 380

Honorarium from Roche. All remaining authors have declared no conflicts of interest. 381

382

383

17

References 384

[1] Australian Institute of Health and Welfare, National Breast and Ovarian Cancer 385

Centre. Ovarian Cancer in Australia: an overview, 2010. In. Canberra: Australian Institute of 386

Health and Welfare & National Breast and Ovarian Cancer Centre; 2010. 387

[2] Lockwood-Rayermann S. Survivorship issues in ovarian cancer: a review. Oncol Nurs 388

Forum; 2006;33: 553-562. 389

[3] Eisenhauer EA, Vermorken JB, van Glabbeke M. Predictors of response to 390

subsequent chemotherapy in platinum pretreated ovarian cancer: a multivariate analysis of 391

704 patients. Ann Oncol; 1997;8: 963-968. 392

[4] Friedlander M, Butow P, Stockler M, Gainford C, Martyn J, Oza A, Donovan HS, 393

Miller B, King M. Symptom control in patients with recurrent ovarian cancer: measuring the 394

benefit of palliative chemotherapy in women with platinum refractory/resistant ovarian 395

cancer. Int J Gynecol Cancer; 2009;19 Suppl 2: S44-S48. 396

[5] Colombo N, Gore M. Treatment of recurrent ovarian cancer relapsing 6-12 months 397

post platinum-based chemotherapy. Crit Rev Oncol Hematol; 2007;64: 129-138. 398

[6] Blackledge G, Lawton F, Redman C, Kelly K. Response of patients in phase II studies 399

of chemotherapy in ovarian cancer: implications for patient treatment and the design of phase 400

II trials. Br J Cancer; 1989;59: 650-653. 401

[7] Emanuel EJ, Young-Xu Y, Levinsky NG, Gazelle G, Saynina O, Ash AS. 402

Chemotherapy use among Medicare beneficiaries at the end of life. Ann Intern Med; 403

2003;138: 639-643. 404

[8] Goff BA, Thompson T, Greer BE, Jacobs A, Storer B. Treatment of recurrent 405

platinum resistant ovarian or peritoneal cancer with gemcitabine and doxorubicin: A phase 406

I/II trial of the Puget Sound Oncology Consortium (PSOC 1602). Am J Obstet Gynecol; 407

2003;188: 1556-1562. 408

18

[9] Gordon AN, Fleagle JT, Guthrie D, Parkin DE, Gore ME, Lacave AJ. Recurrent 409

epithelial ovarian carcinoma: a randomized phase III study of pegylated liposomal 410

doxorubicin versus topotecan. J Clin Oncol; 2001;19: 3312-3322. 411

[10] Beesley V, Clavarino A, Webb P, Wyld D, Francesconi A, Horwood K, Doecke J, 412

Loos C, Green A. Ranked importance of outcomes of first-line and repeated chemotherapy 413

among ovarian cancer patients. Support Care Cancer; 2010;18: 943-949. 414

[11] Merritt MA, Green AlC, Nagle CM, Webb PM. Talcum powder, chronic pelvic 415

inflammation and NSAIDs in relation to risk of epithelial ovarian cancer. Int J Cancer; 416

2008;122: 170-176. 417

[12] Tothill RW, Tinker AV, George J, Brown R, Fox SB, Lade S, Johnson DS, Trivett 418

MK, Etemadmoghadam D, Locandro B, Traficante N, Fereday S, Hung JA, Chiew Y-E, 419

Haviv I, Gertig D, DeFazio A, Bowtell DDL. Novel molecular subtypes of serous and 420

endometrioid ovarian cancer linked to clinical outcome. Clin Cancer Res; 2008;14: 5198-421

5208. 422

[13] Jordan SJ, Green AlC, Whiteman DC, Moore SP, Bain CJ, Gertig DM, Webb PM. 423

Serous ovarian, fallopian tube and primary peritoneal cancers: a comparative epidemiological 424

analysis. Int J Cancer; 2008;122: 1598-1603. 425

[14] Beesley V, Price M, Butow P, Green A, Olsen C, Webb P, Australian Ovarian Cancer 426

Study, AOCS-QoL Study Investigators. Physical activity in women with ovarian cancer and 427

its association with decreased distress and improved quality of life. Psychooncology; 428

2011;20: 1161-9. 429

[15] Pectasides D, Psyrri A, Pectasides M, Economopoulos T. Optimal therapy for 430

platinum-resistant recurrent ovarian cancer: doxorubicin, gemcitabine or topotecan? Expert 431

Opin Pharmacother; 2006;7: 975-987. 432

19

[16] Rocconi RP, Case AS, Straughn JM, Jr., Estes JM, Partridge EE. Role of 433

chemotherapy for patients with recurrent platinum-resistant advanced epithelial ovarian 434

cancer: A cost-effectiveness analysis. Cancer; 2006;107: 536-543. 435

[17] Basen-Engquist K, Bodurka-Bevers D, Fitzgerald M, Webster K, Cella D, Hu S, 436

Gershenson D. Reliability and validity of the Functional Assessment of Cancer Therapy-437

Ovarian. J Clin Oncol; 2001;19: 1809-1817. 438

[18] Ibrahim JG, Molenberghs G. Missing data methods in longitudinal studies: a review. 439

Test (Madrid, Spain); 2009;18: 1-43. 440

[19] Sloan JA, Cella D, Hays RD. Clinical significance of patient-reported questionnaire 441

data: another step toward consensus. J Clin Epidemiol; 2005;58: 1217-1219. 442

[20] Courneya KS, Karvinen KH, Campbell KL, Pearcey RG, Dundas G, Capstick V, 443

Tonkin KS. Associations among exercise, body weight, and quality of life in a population-444

based sample of endometrial cancer survivors. Gynecol Oncol; 2005;97: 422-430. 445

[21] Alsop K, Fereday S, Meldrum C, deFazio A, Emmanuel C, George J, Dobrovic A, 446

Birrer M, Webb P, Stewart C, Friedlander M, Fox S, Bowtell D, Mitchell G. BRCA mutation 447

frequency and patterns of treatment response in BRCA mutation-positive women with 448

ovarian cancer: a report from the Australian Ovarian Cancer Study Group. J Clin Oncol; 449

2012;30: 2654-2663. 450

[22] Gordon L, Scuffham P, Beesley V, Green A, DeFazio A, Wyld D, Clavarino A, 451

Australian Ovarian Cancer Study, Webb P. Outcomes and related medical costs for women 452

with ovarian cancer in Australia: a patient-level analysis over 2½ years. Int J Gynecol 453

Cancer; 2010;20: 757-65. 454

[23] Shih Y-CT, Halpern MT. Economic evaluations of medical care interventions for 455

cancer patients: how, why, and what does it mean? CA Cancer J Clin; 2008;58: 231-244. 456

457

20

Tables and Figures 458

Table 1: Demographic and clinical characteristics of all participants and by missingness type 459

Figure 1. Flow of participant recruitment and data contribution to this analysis 460

Figure 2. Mean quality of life score by dropout time stratified by platinum status 461

Figure 3. Chemotherapy regimen administered after first progression, by platinum status 462

Figure 4a. Least squares mean quality of life scores following second-line chemotherapy 463

in women with ovarian cancer treated for platinum-sensitive (n=128) versus -resistant 464

(n=44) disease. Main model: including all available data. 465

Figure 4b. Least squares mean quality of life scores following second-line chemotherapy in 466

women with ovarian cancer treated for platinum-sensitive (n=112) versus -resistant (n=29) 467

disease. Sensitivity model: not including women who dropped out due to death or incapacity. 468

Figure 4c. Least squares mean quality of life scores following second-line chemotherapy in 469

women with ovarian cancer treated for platinum-sensitive (n=97) versus -resistant (n=31) 470

disease. Sensitivity model: not including women with baseline only data. 471

Figure 5. Wellbeing subscale breakdown of overall quality of life presented in figure 4a. 472 473

21

Table 1: Demographic and clinical characteristics of all participants and by missingness 474 type 475 476 All

participants (n=172)

Participants with complete data (n=104)

Dropout due to death or incapacity

(n=31)

Dropout due to completed

study or unknown

reason (n=37) Platinum sensitive, % 74 77 52* 86

Age at progression, mean (SD)

60 (10) 60 (9) 62 (10) 58 (12)

Had partner, % 71 74 53* 76

Stage III or IV at diagnosis, %

96 96 97 95

QoL at second-line chemotherapy initiation, mean (SD)

105 (22) 108 (22) 96 (22)* 105 (21)

Survival from second-line chemotherapy initiation (months), median

26 30 9 * 29

* Different from all participants 477 478

22

479 480

481

482 483 484 485 486 487 488 489 490 491 492 493

494

495

496

497

498

499

500

501

Figure 1. Flow of participant recruitment and data contribution to this analysis 502 a Completed follow-up for parent study 503

PROSPECT study April 2003 and January 2005 N = 122

PROSPECT study April 2003 and January 2005

N = 122

AOCS-QOL study May 2005 and March 2007

N = 798

Completed questionnaire within 2 months of starting chemotherapy for first progression

N = 52

Completed questionnaire within 2 months of starting chemotherapy for first progression

N = 130

Platinum status was classifiable i.e. data used in this analysis N = 172 Platinum-sensitive (PS) N = 128, 74%

Platinum-resistant (PR) N = 44, 26%

3 month follow-up data = 128 (74%) • PS withdrawals: death = 7, incapacity = 4, completed follow-upa = 7, unknown = 13. • PR withdrawals: death = 4, incapacity = 5, unknown = 4.

6 month follow-up data = 104 (60%) • PS withdrawals: death = 2, incapacity = 3, completed follow-upa = 4, unknown = 8. • PR withdrawals: death = 4, incapacity = 2, unknown = 1.

23

Platinum-sensitive Platinum-resistant 504

505 Figure 2. Mean quality of life score by dropout time stratified by platinum status 506

95

100

105

110

115

120

2nd-linechemotherapy

start

3 months 6 months

Qua

lity

of li

fe sc

ale

(FAC

T-0)

rang

ing

from

0-1

48

Complete data n=80

6 month dropout n=17

3 month dropout n=31

95

100

105

110

115

120

2nd-linechemotherapy

start

3 months 6 months

Complete data n=24

6 month dropout n=7

3 month dropout n=13

24

507

508

509

510

511

512

513

514

515

516

517

518

519

Figure 3. Chemotherapy regimen administered after first progression, by platinum 520

status 521

45%

26%

16%

10%

1% 2%

7%

18%

11%

43%

11% 9%

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

Plat

in a

ndta

xane

age

nts

Plat

in &

oth

erag

ent(

s)

Plat

in o

nly

Lipo

som

aldo

xoru

bici

n

Taxa

ne o

nly

Oth

er

% re

ceiv

ed re

gim

en

Platinum-sensitive Platinum-resistant

25

Figure 4a. Least squares mean quality of life scores following second-line chemotherapy in women with ovarian cancer treated for platinum-sensitive (n=128) versus -resistant (n=44) disease. Main model: including all available data.

Figure 4b. Least squares mean quality of life scores following second-line chemotherapy in women with ovarian cancer treated for platinum-sensitive (n=112) versus -resistant (n=29) disease. Sensitivity model: not including women who dropped out due to death or incapacity. Figure 4c. Least squares mean quality of life scores following second-line chemotherapy in women with ovarian cancer treated for platinum-sensitive (n=97) versus -resistant (n=31) disease. Sensitivity model: not including women with baseline only data. Note: All models adjusted for age at progression and whether currently on chemotherapy. Marital status, education, second-line chemotherapy with platinum or liposomal doxorubicin were tested and not included as covariates due to their lack of association. Note: A difference of 8 on the FACT-O scale is considered clinically significant

105

111

115

107

105 103

90

100

110

120

2nd-line chemotherapystart

3 months 6 monthsQua

lity

of li

fe ra

ngin

g fr

om 0

-148

107

113 116

112

111

108

90

100

110

120

2nd-line chemotherapystart

3 months 6 monthsQua

lity

of li

fe ra

ngin

g fr

om 0

-148

106

112 115

109

107

104

90

100

110

120

2nd-line chemotherapystart

3 months 6 monthsQua

lity

of li

fe ra

ngin

g fr

om 0

-148

platinum-sensitive platinum-resistant

Change in platinum-sensitive group p<0.001 Change in platinum-resistant group p=0.229

Change in platinum-sensitive group p<0.001 Change in platinum-resistant group p= 0.260

Change in platinum-sensitive group p<0.001 Change in platinum-resistant group p= 0.167

26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Figure 5. Wellbeing subscale breakdown of overall quality of life presented in figure 4a. 37 38 Note: A difference of 2 on the wellbeing scales and 2.5 on the ovarian cancer-specific concerns scale is considered to be a 39 clinically significant least squares mean difference. 40 41 Statistical support (p<0.05) for changes were noted in physical, functional, emotional wellbeing and ovarian cancer-specific 42 concerns among women who had platinum-sensitive disease. 43 44

18.7

20.1

22.4

20.5

19.1 18.5

15

17

19

21

23

25

2nd-linechemotherapy

start

3 months 6 months

Phys

ical

wel

lbei

ng 0

-28

22.1 22.5 22.3

22.1 22.4 22.2

15

17

19

21

23

25

2nd-linechemotherapy

start

3 months 6 monthsSoci

al/f

amily

wel

lbei

ng 0

-28

17.2 18.2 18.3

16.6 17.2 17.7 15

17

19

21

23

25

2nd-linechemotherapy

start

3 months 6 months

Emot

iona

l wel

lbei

ng 0

-24

17.6

19.1 19.6 18.8

18.2 17.4 15

17

19

21

23

25

2nd-linechemotherapy

start

3 months 6 months

Func

tiona

l wel

lbei

ng 0

-28

29.5 31.1 32.1

28.7 28.1 27.5

22

25

28

31

34

37

2nd-linechemotherapy

start

3 months 6 months

Ova

rian

canc

er-s

peci

fic c

once

rns

0-40

platinum-sensitive platinum-resistant