Relapsed disease and aspects of undetectable MRD and treatment discontinuation

A 69-year-old female patient presented 2 years ago with progressive chronic lymphocytic leukemia (CLL) into Binet stage C and Rai III after she had undergone a 5-year period of watch and wait. Her CLL risk profile showed a mutated immunoglobulin heavy chain variable region genes (IGHV) status, no TP53 mutation, and no del(17p); only a del(13q) was detected by fluorescence in situ hybridization (FISH). Her CLL International Prognostic Index (CLLIPI) was 2 with intermediate risk. Because of her favorable CLL risk profile, her minor comorbidity burden (diabetes mellitus type II and hypothyroidism), and her age, treatment with bendamustine and rituximab (BR) was started. After 1 cycle of BR, the patient developed severe myelotoxicity (transfusion-dependent anemia without signs of hemolysis and prolonged neutropenia) and severe neurotoxicity. Treatment with BR was stopped. Because lymphocytes and lymph nodes had been reduced by nearly 50% at this point, no other treatment was initiated. The patient again underwent a watch and wait strategy. Her red blood cells recovered; however, 8 months later, her lymph nodes were symptomatically enlarged again, her lymphocytes were up to 35.5 g/L, hemoglobin was at 111 g/L, and platelet count was within the normal range. Because of the persisting neurotoxicity, frequent visits to the outpatient clinic were difficult to manage for her, and therefore, treatment with ibrutinib 420 mg daily was chosen for relapse therapy. Ibrutinib was well tolerated, with some mild rash at the beginning and a few episodes of epistaxis. Eighteen months after treatment initiation, lymph node enlargement had disappeared, and hemoglobin and platelet count were within normal range. Lymphocytosis is persisting with 19 g/L. Because neurotoxicity has also recovered, the patient has an excellent quality of life.

Relapsed CLL does not necessarily require immediate retreatment of patients, particularly when they are asymptomatic. Patients should meet International Workshop on CLL (iwCLL) treatment indication for relapse treatment.¹  When CLL reoccurs after frontline therapy, patients may still undergo a watch and wait strategy depending on the dynamics of relapse and prior therapy (relapse after treatment-free interval vs progression during ongoing treatment). Patients progressing while being on continuous therapy with novel B-cell receptor inhibitors (BCRis; including Bruton’s tyrosine kinase [Btk] inhibitors as well as phosphatidylinositol 3-kinase [PI3K] inhibitors) should not stop the respective treatment before an alternative therapy has been initiated or at least planned within the next days, because rebound phenomena with rapidly progressing disease have been observed in this situation.

For the selection of an optimal second-line or later-line therapy, the intensity and side effects of previous therapies as well as the duration of the last response in addition to the CLL risk profile should be considered. A reassessment of genetic defects should also be performed to identify genetic evolution, such as acquisition of TP53 aberration by FISH, and in addition, it should be reassessed by molecular sequencing (ie, Sanger technique or next generation sequencing), particularly if patients are of younger age and fit enough for cellular therapy. Complex karyotype is not yet being tested by routine, because published data do not show consistent results.^2,3  For some cohorts, the poorer prognosis of CLL carrying complex karyotype can be explained by an association with TP53 aberration.⁴  For patients progressing on Btk inhibitor therapy, Btk mutations or PLCγ mutations are so far not routinely tested.⁵  Similarly, for patients progressing on bcl2 inhibitor therapy with venetoclax, no molecular genetic testing is recommended, although Gly101Val mutations in bcl2 may be detected in up to 75% of high-risk patients refractory to venetoclax.^6,7  However, early relapses and progressing disease on continuous therapy are frequently associated with an acquisition of additional genetic aberrations associated with a poor prognosis or expansion of prior existing subclones.⁸ 

As outlined above, early relapses after chemoimmunotherapy (CIT) are associated with an enrichment of adverse genetic risk factors, such as a genetic TP53 aberration, resulting from selection pressure and DNA damage under chemotherapy.⁹  For the interpretation of clinical trial data, it is important to note that patients being included in relapse trials are in general more heterogenous than those being included in frontline trials. The outcome of these trials is particularly depending on percentage of included CLL cases carrying adverse genetic risk factors.

With the exception of the combination of the bcl2 inhibitor venetoclax plus rituximab, none of the chemotherapy-free regimens containing novel targeted agents have been tested against CIT in the relapsed/refractory setting. However, the excellent outcome of the novel inhibitor treatment vs momontherapy with CD20 antibody within phase 3 trials and their remarkable efficacy in CLL with high-risk aberrations, such as TP53 mutation or deletion, have defined these agents as treatment of choice in relapsed CLL.

Ibrutinib is the inhibitor with the most evidence in relapsed CLL. After a 5-year observation time of the first phase 2 trial, median progression-free survival (PFS) is 51 months in pretreated CLL patients (Table 1).¹⁰  A phase 3 trial comparing ibrutinib with ofatumumab in relapsed patients demonstrated superior efficacy for ibrutinib with respect to response, PFS, and overall survival (OS). After a median observation time of 44 months, PFS at 3 years was 59% with ibrutinib in comparison with 3% with ofatumumab.³  Within a registry, median PFS with ibrutinib monotherapy was lower with 35 months for relapsed CLL.¹¹  The lower PFS rate is likely because of a higher rate of treatment interruptions outside of clinical trials (up to 42%), which were mostly done because of toxicities.¹¹  Ibrutinib combined with bendamustine and rituximab (BRi) was also superior to BR alone.¹²  However, the contribution of BR to this superiority of BRi is not clear, because a third arm with ibrutinib alone was missing. Because PFS rates of ibrutinib monotherapy from other trials seem similar to those of BRi, this treatment regimen has not become an established treatment option.

Other Btk inhibitors will be shortly approved in relapsed CLL. In vitro acalabrutinib showed a greater relative selectivity than ibrutinib for BTK.¹³  A recently presented phase 3 trial in 310 relapsed CLL patients compared acalabrutinib with investigator’s choice of either BR or idelalisib plus rituximab. PFS rates after 1 year with acalabrutinib were 88% vs 68% with BR or idelalisib plus rituximab.¹⁴  With respect to the side effect profile of acalabrutinib, which is possibly slightly different from ibrutinib, data from head to head comparison are still pending. However, because acalabrutinib targets the same site at Btk, it is not appropriate for CLL cells carrying Btk resistance mutation.

The PI3K inhibitor idelalisib combined with rituximab (Idela + R) was compared with rituximab plus placebo in another phase 3 trial in relapsed patients (including 42% with a TP53 aberration) (Table 1).¹⁵  The combination was also found to be significantly better with respect to PFS (median PFS for Idela + R of 20.3 months) and OS (median of 40.6 months) after longer follow-up.¹⁶  However, because of toxicities, particularly diarrhea, cytomegalovirus reactivations, and Pneumocystis pneumonia, treatment interruptions and discontinuations were not uncommon with this regimen, resulting in shorter duration of response.¹⁷  However, particularly in CLL with unfavorable TP53 mutation or deletion, response durations were remarkable.¹⁸  Another recently approved PI3K inhibitor is duvelisib, which demonstrated its efficacy in comparison with ofatumumab within a phase 3 study.¹⁹  Duvelisib showed significantly prolonged PFS compared with ofatumumab (13.3 vs 9.9 months, hazard ratio = 0.52, P < .0001) (Table 1). The toxicity profile in comparison with idelalisib is possibly more favorable, but data from a head to head comparison are missing.

The BCL2 inhibitor venetoclax has demonstrated remarkable single-agent activity in CLL. A phase 2 study in 158 mostly relapsed/refractory CLL patients with 17p deletions demonstrated an overall response rate of 77%, including a 20% complete response rate (CRR).²⁰  The 2-year PFS was 54% (95% confidence interval, 45%-62%) (Table 1).²¹  Forty-eight (30%) of 158 patients achieved a minimal residual disease (MRD) below 10⁻⁴ at least once in the course of the study.

The combination of venetoclax with rituximab (VR) is approved as the first fixed duration treatment regimen with novel agents based on the data from the MURANO study.²²  In total, 389 CLL patients were randomized between 6 cycles of BR or 24 months of venetoclax plus rituximab. The estimated 2-year PFS was 85% for VR and 36% for BR. VR also produced a significantly prolonged OS. With an extended median follow-up of 36 months, 130 (67%) of 194 patients completed the 2-year treatment, and with a median observation time of 9.9 months after completion of treatment with VR, only 16 of 130 patients (12%) showed disease progression.²³ 

The role of CIT has certainly decreased in relapsed CLL with novel agents being available. However, in some situations—very favorable genetic profile, very long-lasting remission to frontline therapy, and questionable patient compliance for oral drug intake—CIT might still be considered, particularly when access to targeted agents is difficult. A recently published European observational study on BR in first relapse of CLL showed a surprisingly long time to next treatment of 31 months for this CIT regimen.²⁴ 

With several options in relapse therapy being available, cellular therapies are nowadays considered as later lines of therapy. According to actual recommendations, allogeneic stem cell transplantation (allo SCTx) in relapsed CLL after CIT should only be considered in high-risk CLL harboring del(17p) or TP53 mutation and if the patient has an individual low transplantation-associated risk or if the patient is refractory to 1 or even 2 novel agents.^25,26 

The prognosis of CLL patients progressing on treatment with BCRi has been very poor in the past.²⁷  Although those patients progressing with Richter’s transformation had the poorest outcome, with <4 months OS, progression of CLL on BCRi therapy resulted in 17.4 months OS only.

Registry data showed that changing treatment to venetoclax on progression during or intolerance to BCRi treatment resulted in a better outcome than changing to alternate BCRi or CIT.²⁸  A phase 2 trial has confirmed this observation in 127 patients who received venetoclax after BCRi failure with a median PFS of 24.7 months for venetoclax treatment after ibrutinib therapy.²⁹  Another cohort of the trial received venetoclax after prior idelalisib failure and showed an estimated 1-year PFS of 79%.³⁰ 

A registry analysis assessed outcomes in patients with prior BCRi therapy with regard to the reason of BCRi treatment discontinuation.²⁸  The main reason for BCRi discontinuation was toxicity (51%) followed by disease progression (29%) and Richter’s transformation. Patients who were retreated because of intolerance of the previously used BCRi had significantly better outcomes than patients who showed disease progression during BCRi treatment.²⁸  This finding is particularly relevant with the upcoming approval of other Btk inhibitors than ibrutinib. The Btk inhibitors acalabrutinib or zanubrutinib might harbor a more favorable toxicity profile than ibrutinib.^31,32  Thirty-three patients who discontinued ibrutinib treatment because of toxicities were treated with acalabrutinib and showed an ORR of 95%. Median PFS has not been reached after a median follow-up of 14.3 months.³¹  During this time, only 9% of the ibrutinib-intolerant patients discontinued treatment with acalabrutinib because of adverse events.³³  Hence, because of market access of acalabrutinib in the United States for Waldenstroem’s lymphoma and some early access programs in Europe, patients can already be switched to an alternate Btk inhibitor when intolerance is the reason for ibrutinib discontinuation.

Subsequent treatments for patients failing venetoclax treatment have not yet been extensively analyzed. In a recently published analysis, 204 venetoclax-treated patients were evaluated, of whom about 47% discontinued treatment because of progressive disease, 21% discontinued treatment because of Richter’s transformation, and 11% discontinued treatment because of mostly hematological adverse events.²⁸  Nineteen patients who were subsequently treated with BCRi showed a good ORR of 69% and a median follow-up that was not reached after 7 months of median follow-up. First data from patients progressing on the MURANO trial who were subsequently treated with ibrutinib indicate as well that BCRi can successfully be used after venetoclax.³⁴  Of 8 patients who were included in this analysis, 7 responded to ibrutinib (6 PR and 1 CR).

Ongoing trials evaluate if retreatment with tailored venetoclax-based combinations is possible and efficacious. First data from a phase 1b trial show that re-exposure to venetoclax induced durable responses in a few patients.³⁵ 

With respect to the limited treatment options after progression on BCRi therapy, those patients have now been classified as high-risk CLL in relapsed situation, and they should undergo allo SCTx if a donor is available and as soon as they have achieved remission with alternate inhibitor therapy.²⁵ 

In summary, the choice of relapse treatment depends primarily on prior treatment, reason for stopping prior treatment, and duration of response. In addition, genetic risk factors play a role in when CIT is considered or patients are considered for allo SCTx or other cellular therapies.²⁵ 

Patients with prior CIT should preferentially switch to inhibitor therapy. Based on the patients’ comorbidity profile and preference between ongoing therapy vs fixed duration therapy, the most suitable treatment, ibrutinib (indefinite therapy) vs venetoclax plus rituximab (24-months fixed duration therapy), should be chosen. Although cardiac comorbidities have to be considered for ibrutinib therapy, sufficient creatinine clearance and willingness to follow the dose escalation scheme, including regular control of tumor lysis syndrome parameters, are necessary for treatment initiation with venetoclax.

If patients had a very long-lasting remission, unfavorable genetic aberrations are excluded, and if the patient wishes a short therapy, CIT can be considered. The CIT of choice would be BR, because later-line therapy with fludarabine plus cyclophosphamide plus R (FCR) is associated with significant toxicities, and BR has been shown to be superior to chlorambucil plus rituximab.³⁶ 

In patients with inhibitor therapy, the reason for stopping treatment should drive subsequent treatment. Patients becoming intolerant to BCRi can be switched to an alternate BCRi, particularly when alternate Btk inhibitors are available. Patients progressing on ongoing BCRi therapy should switch to venetoclax alone or in combination with rituximab and vice versa for patients progressing on ongoing venetoclax therapy.

Other BTK inhibitors are currently being developed to overcome the development of resistance mutations. The substance ARQ 531 has shown promising efficacy in the context of BTK C481S mutations.³⁷ 

If prior fixed duration treatment of venetoclax plus CD20 antibody (for example, venetoclax plus obinutuzumab in frontline treatment, which has been recently approved by the Food and Drug Administration) or venetoclax alone was administered and the patient had a treatment-free interval of a couple of years, re-exposure to bcl2 inhibitor is a possible treatment option. However, data are so far very limited, and therefore, strict recommendation regarding the length of a required treatment-free interval cannot be given.

All of these aspects emphasize that careful evaluation of CLL at the time of relapse, including genetic retesting, is warranted and that prospective clinical trials, including long-term follow-ups, are urgently needed to establish an optimal sequencing strategy.

A 61-year-old patient who was diagnosed with CLL 20 years ago received prior treatment with intermittent chlorambucil therapy in a private outpatient clinic. After 7 years, chlorambucil monotherapy treatment was changed to 6 cycles of chlorambucil plus rituximab owing to nonresponse. After a few months of treatment-free interval, CLL progressed again, and the patient presented the first time at our institution with anemia and thrombocytopenia because of bone marrow (BM) infiltration as well as an extensive lymphadenopathy in all areas. del(17p) was detected by FISH. The patient was offered treatment within a clinical trial (the CLL2BAG study of the German CLL study group [GCLLSG]³⁸ ) evaluating MRD-guided treatment with venetoclax plus obinutuzumab after an optional bendamustine debulking. He received 2 cycles of bendamustine within the clinical trial only for debulking, because chemotherapy is otherwise not an adequate treatment of CLL carrying TP53 dysfunction. Debulking was followed by 6 cycles of induction treatment with venetoclax plus obinutuzumab followed by a maintenance therapy with continuous venetoclax and 3-monthly obinutuzumab infusions. After 15 months in total, treatment was stopped according to the protocol, because no MRD was detected at 2 follow-up visits in peripheral blood (pB) by flow. At final restaging, he was assessed as PR with undetectable MRD (uMRD) because of a residual abdominal lymph node of 1.8 cm. The patient presented 1 year later with increasing lymph nodes and mild lymphocytosis. Interestingly, no del(17p) or TP53 mutation was detected in pB. The patient was offered participation in another trial evaluating a combination therapy of idelalisib, the Btk inhibitor tirabrutinib, and obinutuzumab and is currently in remission again.

MRD in CLL is defined as the number of leukemic cells that can be detected by different techniques in pB or BM after treatment. uMRD is currently defined as the presence of <1 CLL cell in 10 000 leukocytes (<10⁻⁴) by either flow or polymerase chain reaction (PCR).¹  After the first reports of deep responses with uMRD in the context of FCR treatment, the significance of MRD in CLL has constantly grown.^39,40  Randomized, controlled trials have demonstrated that MRD status after induction treatment is an independent predictor of survival and PFS in addition to genetic risk factors, such as IGHV status, which determine regrowth kinetics^41,42  (Table 2). In 2016, the European Medicines Agency allowed the use of uMRD as an intermediate end point in randomized trials that are used for drug approval.

MRD detection is performed using pB or BM aspirate. The site of MRD detection is relevant, because compartment effects have been observed, with some substances being more effective in pB than in the BM or lymph nodes. The established MRD detection level of <10⁻⁴ in pB can currently be achieved by 3 different methods: multicolor flow cytometry (FCM), real-time quantitative polymerase chain reaction (RQ-PCR), and high-throughput sequencing (HTS), each with their specific advantages (Table 3).

The European Research Initiative on CLL has established a harmonized multicolor FCM protocol that has been widely implemented and led to highly comparable MRD results.⁴³  In contrast to PCR-based methods, this assay does not need pretreatment samples and can be adapted by most laboratories using cytometers with 6 or more colors. Using this simple and reliable method, a detection limit of 10⁻⁵ can be achieved.

A detection limit <10⁻⁵ has also been reported for MRD assessment methods using RQ-PCR with patient-specific primers.⁴⁴  In this method, the disease-specific IGHV is sequenced, and allele-specific oligonucleotide probes are designed to expand the malignant IGHV gene during polymerase chain reaction (PCR). Hence, the assay uses patient-specific primers, and real-time quantitative (RQ)-PCR methodology is quantitative. Its main advantage compared with FCM is that it can be performed on less material and nonfresh samples, including frozen samples. In contrast to RQ-PCR approaches, digital droplet PCR does not require preparation of a disease-specific reference standard curve.

Another DNA-based method of MRD detection that does not require patient-specific primers is HTS of immunoglobulin heavy chain (IGH) using consensus primers for amplification of all IGH genes followed by HTS for quantification of the disease-specific IGH sequence.⁴⁵  With this approach, an MRD detection limit of 10⁻⁶ is reported to be achievable.

In the CIT era, the FCR regimen yielded the highest rates of uMRD, with up to 50% in pB by ITT analysis. Although new inhibitor therapy is more efficacious than CIT, there is a wide range of MRD response rates among them.

The Btk inhibitor ibrutinib excellently controls CLL, but it is more active in the lymph node compartment. By this, ibrutinib leads to a redistribution of lymphocytes into the pB, and therefore, uMRD is only achieved by a minority of patients. Interestingly, in very high-risk CLL ibrutinib monotherapy, MRD status (low MRD [<10⁻²] vs high MRD [≥10⁻²]) was not associated with a difference in PFS (Table 4).⁴⁶  However, these results could also be owing to the small number of patients being analyzed. In contrast to that, the combination of ibrutinib and BR resulted in a pB uMRD rate (by FCM) at 36 months of 26.3% and a higher PFS rate in uMRD patients compared with patients with an MRD ≥ 10⁻⁴ (88.6% vs 60.1%).¹²  A phase 2 trial for pretreated and untreated CLL patients evaluated the time-limited combination treatment of ibrutinib and obinutuzumab, which led to pB uMRD in 44% (Table 4).⁴⁷  Treatment was stopped in patients without detectable MRD in pB in 2 consecutive measurements. Longer follow-up of this trial and other ongoing trials will show if planned treatment discontinuation is feasible and will produce long-lasting treatment-free intervals. Similarly, the combination of FCR with ibrutinib in frontline therapy for fit patients results in impressive uMRD rates. Whether this combination achieves longer treatment-free intervals than those historically reported with FCR alone will have to be shown in the future.

No MRD data have yet been reported for the pivotal studies of the PI3K inhibitor idelalisib in combination with rituximab or with BR as well as for the newer PI3K inhibitor duvelisib (Table 1).

The Bcl2 inhibitor venetoclax has shown very impressive response rates, including CRs and remarkable rates of uMRD even in patients with very high-risk profiles (Table 4). Response rates with respect to uMRD in pB were even higher when venetoclax was combined with anti-CD20 antibodies.^22,38  The combination of venetoclax and rituximab in relapsed CLL showed uMRD in pB in 62% of all patients 2 to 3 months after the end of combination treatment with uMRD, translating into considerably longer PFS.⁴⁸  In the CLL2-BAG study, an optional debulking with bendamustine was followed by an induction and MRD-guided maintenance treatment of venetoclax and obinutuzumab in patients with treatment-naïve or r/r CLL. Here, a uMRD rate of 83% was achieved. The same drug combination has been investigated for frontline therapy within the phase 3 CLL14 trial. The rate of uMRD in this phase 3 trial after fixed duration of 12 months treatment with venetoclax plus obinutuzumab when including only comorbid patients was 75% in pB and 57% in BM measured by ASO-PCR.⁴⁹  The oral combination treatment of ibrutinib and venetoclax has also shown promising preliminary results. Within the phase 2 CLARITY trial, 23 of 40 (58%) patients with r/r CLL have achieved pB uMRD (<10⁻⁴) after 12 months of combined therapy, of which 13 (33%) have even reached MRD levels of <10⁻⁵.⁵⁰  A similar combination regimen consisting of ibrutinib plus venetoclax administered in frontline therapy reached 61% uMRD in BM measured by flow after 12 months.⁵¹ 

The MRD status of patients undergoing allo SCTx at the 12-month landmark analysis after allo-SCTx highly correlated with long-term clinical outcome (Table 4).⁵²  However, the positive impact of uMRD for long-lasting PFS after other cellular therapies, such as chimeric-antigen-receptor T cells, still has to be demonstrated.

Although first analyses have associated uMRD with increased PFS rates in patients treated with venetoclax-containing regimens,²³  more mature data from large trials are required to confirm the value of MRD status as an independent predictive factor for survival in the context of novel agents.

Outside of clinical trials, there are only a few scenarios in which MRD assessment is indicated or could be at least considered. Highly sensitive assessment of residual disease is particularly vital in treatment settings aiming at a cure of CLL, like allo SCTx and possibly, CAR T-cell therapy. In allo SCTx, continued serial assessments are useful to detect low-level molecular relapse that might influence immunosuppression or maintenance therapy.⁵² 

MRD testing can also be useful in selected patients receiving highly effective combination treatments of novel agents, such as venetoclax plus rituximab, outside of clinical trials to evaluate treatment outcome and allow for prognostic estimations.²³  In contrast, in patients receiving indefinite treatments that rarely lead to uMRD (eg, ibrutinib or idelalisib), MRD assessment should not be routinely done, because it seems to lack predictive value.⁴⁶  Conversely, positive MRD results might worry patients and necessitate additional counseling. According to the recently updated iwCLL guidelines, symptomatic disease remains the main criterion for treatment, and MRD status should not play a role in establishing treatment indications outside of clinical trials and the post–allo-SCTx setting.¹ 

MRD-guided therapy is currently being evaluated in several clinical phase 2 trials. One phase 2 trial, for example, analyzes treatment extension with ibrutinib and obinutuzumab after FC plus obinutuzuamb plus ibrutinib based on the MRD results after 3 cycles of therapy. Other phase 2 studies evaluate stopping of novel agent combinations based on 2 consecutive uMRD testings³⁸  or a combination of uMRD testing and the time to uMRD detection.⁵⁰  However, prospective trials randomizing between stopping and continuing treatment when uMRD has been achieved are still missing. Moreover, there are many additional questions to be addressed:

• Which method of measuring MRD (see above) should be applied and can be reimbursed?

• Which time point is optimal for uMRD testing either with novel agent combination or with novel agents alone?

• Should MRD-positive patients continue with the same treatment, or should they change the substance, particularly when new subclones are detected?

• Should patients with TP53 disruption, who may progress earlier after stopping venetoclax-based therapy,⁴⁹  continue in spite of uMRD testing?

• Because different agents have different efficacy in different compartments (eg, ibrutinib is more efficacious in lymph nodes and spleen, venetoclax is more efficacious in pB and BM, and anti-CD20 antibodies are more efficacious in pB), how should we measure MRD outside pB and BM? Will liquid biopsies play a role?

In the era of CITs, treatment was discontinued simply because of severe toxicities, which made longer treatment impossible. With continuous BCRi therapy, increasing CR and uMRD rates are observed with longer treatment duration. It is still unknown if treatment can be stopped at a certain time point when CR or even uMRD has been achieved. Additionally, although venetoclax-based combination therapies result in high rates of uMRD, the optimal duration of therapy is unknown in this situation as well.

Considering that, in CLL, not only response rates and MRD levels but also, genetic aberrations in subclones could be considered for future treatment approaches, it will be one of the biggest challenges in the next years of CLL therapy to transfer the complexity of these approaches into clinical practice.

Barbara Eichhorst, Department I for Internal Medicine, Center of Integrated Oncology Aachen Bonn Cologne Duesseldorf, University of Cologne, Kerpenerstr. 62, 50924 Cologne, Germany; e-mail: barbara.eichhorst@uk-koeln.de.