Original article. Combined modality treatment for primary gastrointestinal non-hodgkin's lymphoma: The Milan Cancer Institute Experience

Transcription

1 Annals of Oncology 4: , Kluwer Academic Publishers. Printed in the Netherlands. Original article Combined modality treatment for primary gastrointestinal non-hodgkin's lymphoma: The Milan Cancer Institute Experience C. Tondini, 1 R. Giardini, 2 F. Bozzetti, 3 P. Valagussa, 1 A. Santoro, 1 R. Bertulli, 1 M. Balzarotti, 1 A. Rocca, 1 F. Lombardi, 4 A. J. M. Ferreri 1 & G. Bonadonna 1 'Division of Medical Oncology, Department of Cancer Medicine; 2 Division of Pathology; ^Division of Surgical Oncology; ^Division of Radiotherapy, Istituto Naiionale Tumori, Milano, Italy Summary Background: The study analyzes clinical-pathologic features, treatment and outcome of all patients with primary lymphoma of the gastrointestinal tract (GI-NHL) seen during the past two decades at the Milan Cancer Institute. Subjects and methods: Clinical and histopathologjcal data from 135 patients presenting with GI-NHL and disease localized within the abdomen were reviewed. Of these, 114 (%) presented with limited disease (stage I and II), while 21 patients were found to have disease involvement of other abdominal organs (i.e., liver, pancreas, peritoneum) or more than one gastrointestinal site and were therefore classified as stage IV. Seventy-three percent presented with lymphoma in the stomach, 15% in the small intestine and % in the large bowel, while in 5 cases multiple localizations of the gastrointestinal tract were documented. Median age was 50 years, with one-fourth of patients older than 60 years. According to the revised Kiel classification for GI-NHL, 61% of patients presented with pure high-grade lymphoma, % high-grade NHL associated with residual low-grade lymphoma, and 30% had low-grade NHL. Nine percent presented with bulky disease, 5% with elevated LDH and 21% with a Karnofsky performance status (PS) «S80. Results: Laparotomy with radical (108 patients) or palliative (15 patients) intent was performed in all patients who were not deemed at high risk of complication from major surgery. Complete removal of all measurable tumor was feasable in 101 patients, with no difference relative to primary site. Surgical morbidity and mortality were 11% and 2%, respectively. Overall, 83% of patients were treated with chemotherapy. Patients who did not receive systemic chemotherapy included 12 managed with surgery alone and 10 who received only postoperative irradiation mainly because of lowgrade lymphoma with superficial disease. Of patients with limited disease, % achieved complete tumor remission. After a median follow-up of 73 months, 13 of 113 patients have relapsed, mostly (70%) outside the gastrointestinal tract. The actuarial 10-yr. freedom from progression (FTP) and overall survival (OS) were % and %, respectively. Aside from age, no other factor revealed a statistically significant impact on outcome. There was only a trend in favor of low-grade histology (FTP 7% vs. 7%), that failed to reach statistical significance. Of patients with advanced abdominal disease, 48% achieved complete remission with chemotherapy with or without prior surgical debulking. Actuarial 10-yr. FTP and OS were 44% and 42%, respectively. In this subset, tumor burden and LDH levels represented the most important prognostic factors affecting outcome. Conclusions: This retrospective study underscores the good results obtained in a wide and unselected population of patients with limited-stage primary GI-NHL following a combined-modality approach that included surgical debulking and systemic chemotherapy for most patients. Surgery alone can be considered adequate treatment for patients with lowgrade NHL disease that does not infiltrate beyond the submucosa. Patients with advanced GI-NHL show a long-term outcome similar to that of patients with advanced NHL arising outside the gastrointestinal tract. Key words: combined modality, gatrointestinal tract, Non- Hodgkin's lymphoma, MALT lymphoma Introduction Although non-hodgkin's lymphomas (NHL) typically arise in lymph nodes, 50% to 60% of cases involve extranodal sites at the time of clinical presentation [1, 2]. The most frequent extranodal sites are the lymphatic tissue of the gastrointestinal tract and of the Waldeyer's ring, followed by bone marrow, skin, thyroid, testis, lung and bone [2, 3]. Recently, there has been renewed interest in primary gastrointestinal NHL (GI- NHL) because of a steady increase in the incidence of gastric lymphoma over the past decade [4] while both incidence and mortality of gastric cancer have declined [5]. Many publications have retrospectively reviewed clinical presentation and outcome of patients with GI- NHL. However, the conclusions that emerged from these studies often yielded conflicting results in terms of adequate treatment and long-term relapse-free survival. The discrepancies were mainly attributable to a number of factors, namely the limited number of patients included, evaluations undertaken before the

2 832 widespread use of modern staging procedures and histopathologic classification, and lack of homogeneous patient characteristics and treatment modalities [6]. Several controversial issues still characterize the diagnostic and therapeutic approaches to a patient with a morphological diagnosis of GI-NHL. In this paper we review the experience accumulated at the Milan Cancer Institute over the last 20 years in one of the largest series of patients with primary GI-NHL treated and followed in a single Institution. Patients and methods One hundred fifty-seven patients seen at the Milan Cancer Institute between 172 and 1 with histologically proven diagnoses of NHL obtained from the gastrointestinal tract excluding pediatric patients and patients with Burkitt's lymphoma were identified from our Pathology archives. Clinical and histopathological data of 135 patients who presented with primary gastrointestinal lymphoma and with follow-up information were reviewed. Primary GI-NHL was defined as a lymphoma diagnosed in a patient who presented with symptoms localized to the gastrointestinal tract without extension of the disease outside the abdomen. All of the original histology slides were reviewed. Tissue material from each case was fixed either in 20% buffered formalin or Bouin's fluid and embedded in paraffin. The sections were stained with hematoxylin and eosin, Giemsa, PAS-hematoxylin and Gomori's silver impregnation for reticulin fibers. Phenotype studies were carried out on tissue sections using a panel of antibodies with a modified avidin-biotin peroxidase complex technique (Vectastain ABC kit, Vector Laboratories, Burlingame, CA). Immunoglobulin was detected using rabbit polyclonal or monoclonal anti-ig heavy and light chain antibodies(dako, Glostrup, Denmark). Lineage markers were identified using appropriate monoclonal antibodies (CD45, Dako LC, Dako; CD45Ra: F8.ll.13 Serotec, Oxford, UK; CD20: L26, Dako; CD45RO: UCHL1, Dako; CD74:LN1 and CDW75:LN2, Biotest serum Institute, Dreieich, D; CD3 polyclonal, Dako). NHL were classified according to MALT lymphoma classification as defined by Isaacson et al. (7, 8j. Attention was paid to the characteristic features of gastrointestinal mucosa associated lymphoid tissue (MALT) lymphomas, such as the presence of intermediate-size lymphoid cells having dense or irregularly shaped nuclei with small nucleoli and a pale cytoplasm (centrocytoid cells), with occasional striking monocytoid appearance; the occurrence of lymphoepithelial lesions, of basophilic blast cells (usually of the immunoblastic type or of plasma cells type), of polytipic germinal centers which can be colonized by centrocytoid cells []. Cases were then classified into low-grade lymphoma (MALT type, centroblastic-centrocytic, B-cell monocytoid and T-zone), purely high-grade lymphoma (primary high-grade) or high-grade lymphoma with areas of residual low-grade component (secondary high-grade). Staging procedures for all patients included complete physical with thorough examination of the Waldeyer's ring, chest roentgenogram, complete blood count and biochemical profile. Bilateral bone marrow core biopsies and bipedal lymphangiography were performed in almost all patients (8% and 2%, respectively). Upper gastrointestinal contrast study or endoscopy were used to investigate the gastrointestinal tract in % and 57% of patients, respectively. Laparotomy with resection or biopsy of the primary lesion was performed in 123 patients (%), along with wedge liver biopsy in 77 patients (57%) and splenectomy in 18 patients. In the remaining 12 patients diagnostic specimens were obtained by means of multiple endoscopic biopsies. Patients were staged according to Ann Arbor classification as modified by Musshoff for lymphomas of the gastrointestinal tract [10]. Patients were also reclassified according to the recently proposed prognostic index that ranks lymphoma patients in three risk levels according to stage, LDH level, tumor mass and number of extranodal sites [11]. The treatment plan first included surgical resection of the gastric or intestinal tumor for all patients who were not deemed to have a high mortality risk from surgery. Subsequently, eligible patients received combined treatment according to prospective studies being performed at the Milan Cancer Institute. Briefly, for patients with stage I-EI disease treatment programs included four to six cycles of chemotherapy with CVP (cyclophosphamide, vincristine and prednisone), BACOP (bleomycin, adriamycin, cyclophosphamide, vincristine and prednisone) or CHOP (cyclophosphamide, adriamycin, vincristine, prednisone) followed by loco-regional radiotherapy to the gastric stump and para-aortic lymph node regions with doses of Gy [12-14]. For patients with advanced disease, treatment protocols included chemotherapy with CVP alternated with ABP (adriamycin, bleomycin and prednisone) [15], CHOP or ProMACE/ MOPP (mechlorethamine, vincristine, procarbazine and prednisone) (unpublished data). Fifty-four patients not eligible to enter prospective studies were treated outside the research protocols. In these patients, treatment programs included chemotherapy with chlorambucil alone or CVP for low-grade lymphoma. In high-grade lymphoma we administered CHOP, MACOP-B (methotrexate, adriamycin, cyclophosphamide, vincristine and bleomycin) or Pro- MACE/CytaBOM (cyclophosphamide, adriamycin, etoposide, cytarabine, methotrexate, bleomycin, vincristine and prednisone) followed by loco-regional radiotherapy for stage I-n disease as well as to sites of bulky disease. Patients were considered in complete remission when all physical and radiological signs of disease had disappeared for a minimum of four weeks. Patients not in complete remission at the end of the treatment program were considered as treatment failures. After completion of treatment, complete responders were followed every 3 to 4 months for the first two years, and every 6 to 12 months thereafter. The probability of freedom-from-progression (FFP) and of overall survival (OS) were computed by the life-table method starting from the date of surgery for patients subjected to radical excision or from the first cycle of chemotherapy for all other patients. Patients who died of causes unrelated to their lymphomas while in first complete remission were censored from the analysis of FFP. All causes of death were regarded as treatment failures for the purpose of survival analysis. The statistical analysis of observed differences was assessed by the log-rank test [16] and all probability values were two-sided. Results Patient characteristics Ninety-eight patients (73%) had primary lymphoma of the stomach, 20 (15%) of the small intestine, 12 (%) of the large bowel, while 5 presented with multiple localization within the gastrointestinal tract (Table 1). The most common complain at presentation was abdominal pain (70%), associated with dyspepsia in one-third of patients and with weight loss in 20% of patients. Upper gastrointestinal tract bleeding, such as hematemesis or melena, was documented in only 10% of patients with gastric lymphoma, while intestinal bleeding accounted for 25% of patients with lymphoma of the large bowel. Patients with lymphoma of the small intestine presented with perforation or intestinal occlusion in 15% and in 20% of the cases, respectively. In general, a clinically palpable mass was evident only in a minority of cases (3%). The median age was 50 years, with 25% of patients being 60 years of age or older. One hundred fourteen patients (%) were clas-

3 Table 1. Main characteristics in 135 patients. The data are in percent. Number of patients Site of disease Stomach Small intestine Large bowel >1 site Age <60 Sex Male Female Histology Low-grade High-grade Stage I III 112 IV Number of sites <2 >2 Performance status 0- <0 LDH Normal Elevated Bulky disease No Yes Prognostic index <1 2 3 Chemotherapy* Without ADM With ADM Stage I-II Stage IV Total Including only patients who received chemotherapy; ADM - adriamycin. sified as having limited disease (Ann Arbor stage I or II). Of this group, the lymphoma had not spread beyond the visceral wall in two-thirds of patients (stage I) while in one-third the disease also involved regional lymph nodes (stage n) (Table 1). According to the modification of the Ann Arbor staging system as proposed by Musshoff et al. [10] for gastrointestinal lymphoma, stage II disease could be further subdivided into stage E,, i.e. involvement of perivisceral lymph nodes only (%), and stage H 2, i.e. involvement of iuxta-regional lymph node stations, such as celiac or retroperitoneal (10%). Twenty-one patients (16%) were found to have lymphomatous involvement of other abdominal organs or sites (liver, pancreas, abdominal wall, peritoneum) or more than one gastrointestinal site, and were therefore classified as having advanced disease (Ann Arbor stage IV). Elevated serum lactic dehydrogenase (LDH) or bulky lymphoma (> 10 centimeters) was documented in % and 11% of patients 833 with limited disease, respectively, and in % and % of patients presenting with advanced disease. Fortyeight percent of patients with advanced disease had a Karnofsky performance status <80 as opposed to 25% of those with limited disease. Only 10% of patients with limited disease had an unfavorable prognostic index (3*2), as compared to 5% of patients with advanced lymphoma. Ninety-five patients (70%) were classified as having high-grade lymphoma (Table 2). In 13 cases there was evidence of residual low-grade lymphoma around areas of blastic predominance. In 40 cases (30%), the histologic appearance was that of low-grade MALT lymphoma. Overall, there was no difference in the percentage of high-grade lymphoma relative to the site of the primary tumor (stomach 66%, small intestine %, large bowel 6%) nor a correlation between stage at presentation and histologic subgroups (data not shown). Table 2. Histology according to the modified Kiel classification. Low-grade 30 MALT type 21 Diffuse, centroblastic-centrocytic 8 Monocytoid, B-cell <1 PTCL, T-zone <1 High-grade, primary 60 High-grade, secondary 10 MALT: mucosa-associated lymphoid tissue; PTCL: peripheral T-cell lymphoma. Treatment Overall, 123 patients (%) underwent laparotomy. Sixty-nine patients were referred to our center after surgery performed elsewhere. From retrospective analysis, surgery was performed with a radical intent in 108 patients, while in 15 patients the purpose was palliative or bioptic. All measurable tumor could be excised in 101 of 108 patients, with no difference between stomach, small intestine and large bowel. Grossly radical excision was performed in 80% of the low-grade lymphomas of the MALT type and in 77% of the primary and secondary high-grade lymphomas. Gastric surgery included 70 subtotal and 15 total gastrectomies. In the 54 patients who were operated on at the Milan Cancer Institute, surgical morbidity (including post-operative abscess, bleeding, anastomosis dehiscence, and pulmonary embolism) was 11% and mortality 2%. Twelve patients did not undergo laparotomy, either because their disease was considered unresectable or because surgical risk was too high. One hundred twelve patients (83%) were treated with chemotherapy. The majority of them (68%) received an adriamycin-containing regimen, with no difference according to histologic subgroups (Table 1). Patients who did not receive chemotherapy included

4 cases who were managed with surgery alone and 10 patients who received only post-operative radiation therapy. The reason for withholding systemic treatment was the histologic diagnosis of a low-grade lymphoma limited to the superficial layers of the gastric wall and the patient's refusal to undergo drug therapy. Eighty-one patients (60%) received radiation therapy, usually as adjuvant treatment following surgery (10 cases) or chemotherapy (70 cases). An 81-year-old patient with low-grade lymphoma limited to the gastric wall was electively treated with radiation therapy alone due to high surgical risk because of poor performance status. 0 to Outcome Stage I-II-H 2 Of the 114 patients with limited disease, 113 (%) were considered to be in complete remission at the end of treatment (Table 3). Of these, % were already free of disease after surgery, while the remaining 27 patients achieved complete remission during chemotherapy or radiotherapy. Only one patient with high-grade lymphoma never achieved complete remission, developing progressive disease after third-line chemotherapy. After a median follow-up of 73 months, 13 complete responders have relapsed. The actuarial 10-year FFP and OS were % and %, respectively (Table 3, Fig. 1). Relapse occurred in 70% of cases outside the gastrointestinal tract. Aside from age (Fig. 2), no other factor, such as stage (I vs. W. x vs. H 2 ), performance status, primary site, bulky presentation, prognostic index or number of involved sites, showed a statistically significant impact on outcome (Table 4). There was a marked trend towards a better FFP for low-grade versus high-grade lymphomas (7% vs. 7%). The lack of a statistical significance is probably due to the relatively limited number of patients with low-grade lymphoma (Fig. 3). Elevated pre-treatment LDH levels were correlated with higher risk of death but not with a higher probability of relapse from complete response. This retrospective case series does not allow an adequate comparison among the different treatment strategies utilized. However, the outcome of the 23 selected patients with disease limited and superficial who were treated with loco-regional therapy only (surgery +/- radiotherapy, radiotherapy alone in one patient) did not differ from that of patients treated with chemother- Table 3. Overview of patient outcome. Complete remission Surgical Chemo +/ radiotherapy 10-yr. FFP 10-yr. Survival FFP: freedom from progression. Stage MI (%) Stage IV (%) H 1- H 1- H YIABS Fig. I. Outcome of all 114 patients with limited disease (OS: survival; FFP: freedom-from-progression). <60yn YIAKS Fig. 2. Freedom-from-progression of patients with limited disease according to age. apy (FFP % vs. %, OS % vs. %). Finally, in this case series no striking advantage emerged by the inclusion of anthracyclines within the chemotherapy regimen (FFP 0% vs. 83%, OS % vs. 8%), irrespective of histologic subgroup. Stage IV Ten of 21 patients (48%) presenting with advanced disease achieved complete tumor remission with systemic treatment (Table 3) and only two have relapsed. The actuarial FFP and survival at 10 years were 44% and 42%, respectively. Factors that negatively affected both FFP and survival with a statistically significant difference were elevated LDH level (FFP 17% vs. 56% and survival 17% vs. 51%, respectively) and prognostic index score equal to 3 (FFP 20% vs. 73%, survival 17% vs. 78%, respectively). Patients with <2 involved sites showed a marked trend towards a better outcome (FFP % vs. 33% and survival 75% vs. 34%). Histology failed to show a significant impact on outcome in this patient subset (FFP 54 vs. 40%, survival 43 vs. 42%),

5 Table 4. Outcome of patients with limited disease. All patients Stage I m 112 Histology Low-grade High-grade Age <60 >60 PS 0- <0 LDH Normal Elevated Site of involvement Stomach Small intestine Large bowel Bulky disease No Yet Prognostic index Number of sites <2 >2 Treatment Surgery +/- RT Chemotherapy FFP OS po.l p p p Low grod* YEARS Fig. 3. Freedom-from-progression of patients with limited disease according to low vs. high-grade histology. nor did age, performance status, bulky disease, or site of involvement. Discussion This study from the Milan Cancer Institute includes one of the largest case series reported in the medical 835 literature. During the last two decades, 135 patients sought medical attention solely because of abdominal symptoms due to GI-NHL. Most patients (%) proved to have disease limited to a single site within the gastrointestinal tract with or without involvement of adjacent lymph nodes, while a minority of patients were found to have more advanced intra-abdominal disease. Surgical approach was planned in 123 patients, and grossly radical resection of the primary was feasable in the majority (82%) of cases. Most patients (83%) received systemic treatment with chemotherapy, while selected patients (those with superficial lowgrade lymphoma) received only local-regional treatment, that is, surgery with or without post-operative irradiation. Complete tumor remission was obtained in % of patients with limited disease and in 48% of those with advanced disease. Actuarial analysis of survival at 10 years shows a long term success rate of % for patients with limited disease and of 44% for patients with advanced disease. The role of elective surgery for patients with primary or even secondary GI-NHL has always been important both as a diagnostic and a therapeutic tool. Surgery has proved to be curative in a relevant fraction of patients with superficial lymphoma, as opposed to only 30% of cases with deeper infiltration of the visceral layers or with regional nodal involvement [17-26]. Adjuvant treatment following surgical removal of primary gastrointestinal disease has definitely improved the chances of cure for patients with disease spread to draining lymph nodes, with a long-term relapse-free survival of 60% with postoperative radiotherapy and 80% with systemic chemotherapy [-30]. However, adjuvant postoperative treatment for patients with low-grade histology and superficial disease remains controversial. In the present study, the outcome of 23 such patients who did not receive systemic treatment was similar to that of patients who were given chemotherapy, suggesting that in selected cases local treatment alone can in fact be considered adequate treatment. In case series where treatment was primarily local, a number of variables, such as stage, bulky disease, histology and anatomical site of disease could predict for outcome [17,, 26, 27, 31, 32]. In the present series, the main factor affecting the outcome of patients with limited-stage GI-NHL was age at presentation, with patients younger than 60 faring better than older patients. Other factors, such as stage, bulky disease, site of the primary, performance status, number of involved regions or the combination of these factors in a prognostic index [11] failed to exert a statistically significant impact on outcome. This is consistent with more recent studies [22, 33] in which adjuvant chemotherapy was utilized in most patients and in accordance with a recent analysis performed at our Institution on a consecutive series of 183 patients with nodal and extranodal stage I-U intermediate- and high-grade NHL treated with combined modality, where only age appeared to significantly affect relapse-free survival [14].

6 836 Histologic grading has always been found important in predicting outcome in case series where local treatment was the only therapeutic modality [,, 26, 31]. However, several histologic classifications have been used through the years by the numerous groups that have analyzed their case series of patients with GI- NHL. Therefore, at present, it is very difficult to reliably compare outcome of studies that have used different histologic classifications and grading. Moreover, a new subtype of non-hodgkin's lymphoma that develops in mucosal tissues (MALT) has recently been described both in the gastrointestinal tract and in other sites [7, 8]. In the present series, there was a trend towards a better disease-free interval for patients with stage I-II low-grade lymphoma as opposed to those with stage I-II high-grade lymphoma (primary and secondary) (10-yr FFP 7% vs. 7%, p 0.07). Taking into account the high rate of grossly radical surgery that was feasible in this population of patients, these results are consistent with those reported in case series that included chemotherapy in the treatment program [33-35]. The outcome of patients who present with disease more advanced in the abdomen is characterized by a long-term success rate of approximately 40%. This is in accord with long-term results reported for patients with advanced nodal and extranodal non-hodgkin's lymphoma [36]. Moreover, a prognostic index [11] recently developed that ranks patients according to stage, tumor bulk, number of disease sites and serum LDH value was found to accurately predict outcome in this series of patients. Therefore, as suggested by other studies [], advanced GI-NHL appear to behave in the same manner as all other advanced lymphomas with comparable histology and prognostic index. In conclusion, the sequential approach with surgery followed by one of the chemotherapeutic regimens utilized in the treatment of NHL could be considered optimal therapy for limited-stage GI-NHL, especially in instances of nodal involvement. In patients who prove to have low-grade lymphoma limited to the superficial layers of the visceral wall, local treatment alone may be adequate. A conservative approach should be considered for patients with a high surgical risk or for patients who refuse laparotomy, while more clinical experience is required before translating a conservative approach into widespread clinical practice. Finally, treatment of advanced disease should follow the general guidelines that apply to all other types of advanced non-hodgkin's lymphoma. References 1. Otter R, Bieger R, Kluin PM et al. Primary gastrointestinal non-hodgkin lymphoma in a population-based registry. Cancer 18; 60: Gosparodowicz MK, Sutcliffe SB, Brown TC et al. Patterns of disease in localized extranodal lymphomas. J Clin Oncol 1; 5: Paryani S, Hoppe RT, Burke JS et al. Extralymphatic involvement in diffuse non-hodgkin's lymphoma. J Chn Oncol 183; 1: Severson RK, Davis S. Increasing Incidence of Primary Gastric Lymphoma. Cancer 10; 66: Silverberg E, Lubera JA. Cancer statistics, 18, CA 18; 3: Bonadonna G, Valagussa P. Should lymphomas of gastrointestinal tract be treated differently from other disease presentations? Eur J Clin Oncol 1; 22: Isaacson PG. Lymphomas of mucosa-associated lymphoid tissue (MALT). Histopatology 10; 16: Lennert K, Feller AC. Low-grade malignant lymphoma of MALT type. In Histopathology of non-hodgkin's lymphomas. Berlin, Heidelberg, New York, Springer-Verlag, 12; Cogliatti SB, Schmid U, Schumacher U et al. Primary B-cell gastric lymphoma: A clinicopathological study of 145 patients. Gastroenterology 1; 101: Musshoff K. Kliniche stadieneinteilung der nicht-hodgkin lymphome. Strahlentherapie 177; 153: Coiffier B, Gisselbrecht C, Vose JM et al. Prognostic factors in aggressive malignant lymhomas: Descripton and validation of a prognostic index that could identify patients requiring a more intensive therapy. J Clin Oncol 1; : Monfardini S, Banfi A, Bonadonna G et al. Improved five years survival after combined radiotherapy-chemotherapy for stage I-II non-hodgkin's lymphoma. Int J Radiat Oncol Biol Phys 180; 6: Bajetta E, Valagussa P, Bonadonna G et al. Combined modality treatment for stage I-II non-hodgkin's lymphoma: CVP versus BACOP chemotherapy. Int J Radiat One Biol Phys 188; 15: Tondini C, Zanini M, Lombardi F et al. Combined modality treatment with primary CHOP chemotherapy followed by local-regional irradiation in stage I-II histologically aggressive non-hodgkin's lymphoma. J Clin Oncol 13; 11: Bajetta E, Buzzoni R, Viviani S et al. II trattamento dei linfomi non-hodgkin, stadio HI-IV. Argomenti di Oncologia 182; 3: Peto R, Pike, MC, Armitage P et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient.ii. Analysis and examples. Br J Cancer 177; 35: Dworkin B, Lightdale CJ, Weingrad DN et al. Primary gastric lymphoma. A review of 50 cases. Digestive Disease and Sciences 182; 27: Brooks JJ, Enterline HT. Primary gastric lymphomas. A clinicopathologic study of 58 cases with long-term follow-up and literaturereview.cancer 183; 51: Contreary K, Nance FC, Becker WE Primary lymphoma of the gastrointestinal tract. Ann Surg 180; 1: Di Marco A, Altini E, Rizzotti A et al. Primary non-hodgkin's lymphomas of gastrointestinal tract: Analysis of 41 cases. Tumori 10; : Kajanti M, Karkinen-Jaaskelainen M, Rissanen P. Primary gastrointestinal non-hodgkin's lymphoma. Acta Oncol 188; 27: Maor MH, Maddux B, Osborne BM et al. Stages IE and IIE non-hodgkin's Lymphomas of the stomach. Comparison of treatment modalities. Cancer 1; 4: List AF, Greer JP, Cousar JC et al. Non-Hodgkin's Lymphoma of the gastrointestinal tract: An analysis of clinical and pathologic features affecting outcome. J Clin Oncol 188; 6: Shiu MH, Karas M, Nisce L et al. Management of primary gastric lymphoma. Ann Surg 182; 15: Paulson S, Sheehan RG, Stone MJ et al. Large cell lymphomas of the stomach: Improved prognosis with complete resection of all intrinsic gastrointestinal disease. J Clin One 183; 1: Weingrad DN, Decosse JJ, Sherlock P et al. Primary gastroin-

7 Book review Emesis in anti-cancer therapy. Mechanisms and treatment. P. L. R. Andrews, G. J. Sanger (eds). Chapman & Hall Medical, London/Glasgow/New York/Tokyo/ Melbourne/Madras, pp, testinal lymphoma: a 30-year review. Cancer 182; 4: Gospodarowicz MK, Bush RS, Brown TC et al. Curability of gastrointestinal lymphoma with combined surgery and radiation. Int J Radiation Oncology Biol Phys 182; : Bellesi G, Messori A, Alterini R et al. Combined surgery and chemotherapy for the treatment of primary gastrointestinal intermediate-or high-grade non-hodgkin's lymphomas. Br J Cancer 18; 60: Salles G, Herbrecht R, Tilly H et al. Aggressive primary gastrointestinal lymphomas: Review of patients treated with the LNH- Regimen. A Study of the Group d'etude des Lymphomes Agressifs. Am J Med 1; 0: Liang R, Chiu E, Todd D et al. Chemioterapy for early-stage gastrointestinal lymphoma. Cancer Chemiother Pharmacol 1; 27: Green JA, Dawson AA, Lessells AM et al. Prognostic factors in gastrointestinal lymphoma. Clin Oncol 181; 7: Dragosics B, Bauer P, Radaszkiewicz T. Primary gastrointestinal non-hodgkin's lymphomas. A retrospective clinicopathologic study of 150 cases. Cancer 1; 55: Aozasa K, Ueda T, Kurata A et al. Prognostic value of histologic and clinical factors in 56 patients with gastrointestinal lymphomas. Cancer 188; 61: Maor MH, Velasquez WS, Fuller LM et al. Stomach conservation in stage IE and IIE gastric non-hodgkin's lymphoma. J Clin Oncol 10; 8: Steward WP, Harris M, Wagstaff J et al. A prospective study of the treatment of high-grade histology non-hodgkin's lymphoma involving the gastrointestinal tract. Eur J Clin Oncol 1; 21: Armitage JO. The place of third generation regimens in the treatment of adult aggressive non-hodgkin's lymphoma. Ann Oncol 1; 2 (Suppl. 1): Received 23 June 13; accepted 28 July 13. Correspondence to: Dr. Carlo Tondini Division of Medical Oncology Istituto Nazionale Tumori via G. Venezian n Milano Italy Annals of Oncology A: 837, 13. This 256-page textbook on emesis in anti-cancer therapy was edited by P. L. R. Andrews and G. J. Sanger, both of whom have been involved in investigating the physiology of anticancer emesis for many years. Sixteen additional experts in the field have contributed to the 11 chapters on the current level of knowledge and treatment of those most distressing side effects, nausea and vomiting. In the first two chapters the pharmacokinetics and the toxicology of chemotherapeutic agents are discussed and a long list of references on these subjects is provided. These two chapters are a sound reiteration of the basic knowledge about cytostatic agents. The next chapter describes the various methods of nausea measurements, and the reader involved in clinical trials on nausea and emesis will see that there is still no consensus on the best way to assess nausea. The next three chapters focus on the physiology of the peripheral and central pathways of vomiting and the possible mechanisms of emesis induced by cytostatic agents. The tremendous research efforts made in the years since the detection of a new class of antiemetics, the 5-HT-3 antagonists, are very well presented in these ninety pages. The authors admit, however, that many questions about the exact mechanisms of vomiting and nausea can still not be answered. Documented by the sparse existing literature, there is a chapter on the mechanism of radiotherapy-induced emesis, a subject which has been neglected for many years. The authors suggest a common mechanism of serotonin or 'substance P' release in the basis of cytotoxic, radiation or ingested toxin-induced emesis. The next chapter deals with the mechanism of action and the pharmacology of specific antiemetic agents. It gives a good explanation of the antiemetic properties of the most important classes of antiemetic drugs with a comparison of the already commercially available as well as the experimentally used 5-HT-3 antagonists. The last two chapters give very useful recommendations on dose and schedule of commonly used anti-emetic drugs (5-HT-3 antagonists included) for treating acute and delayed nausea and vomiting, and address anticipatory nausea and vomiting and its treatment possibilities. In conclusion, this textbook contains an excellent updated summary of recent progress in the treatment of chemotherapy-induced nausea and vomiting. However, as the chapters were written by many different authors, some overlapping is inevitable. The book can be recommended to medical oncologists who are particularly interested in this field and want to brush up on their knowledge. K. Buser Bern