Select men benefit from androgen deprivation therapy delivered with salvage radiation therapy after prostatectomy.

BACKGROUND: Which men benefit most from adding androgen deprivation therapy (ADT) to salvage radiation therapy (SRT) after prostatectomy has not clearly been defined; therefore, we evaluated the impact of ADT to SRT on failure-free survival (FFS) in men with a rising or persistent PSA after prostatectomy. METHODS: We identified 332 men who received SRT after prostatectomy from 1987 to 2010. Recursive partitioning analysis (RPA) identified favorable, intermediate and unfavorable groups based on the risk of failure after SRT alone. Kaplan-Meier and log-rank tests compared FFS with and without ADT. RESULTS: Forty-three percent received SRT alone and 57% received SRT with ADT (median 6.6 months (interquartile range (IQR) 5.8-18.1) ADT). Median SRT dose was 70 Gy (IQR 70-70), and median follow-up after SRT was 6.7 years (IQR 4.5-10.8). On Cox's proportional hazard regression, ADT improved FFS (adjusted hazard ratio 0.60, 95% confidence interval: 0.42-0.86; P=0.006). RPA classified unfavorable disease as negative surgical margins (SMs) and preradiation PSA of ⩾0.5 ng ml-1. Favorable disease had neither adverse factor, and intermediate disease had one adverse factor. The addition of ADT to SRT improved 5-year FFS for men with unfavorable disease (70.3% vs 23.4%; P<0.001) and intermediate disease (69.8% vs 48.0%; P=0.003), but not for men with favorable disease (81.2% vs 78.0%; P=0.971). CONCLUSIONS: The addition of ADT to SRT appears to improve FFS for men with a preradiation PSA of ⩾0.5 ng ml-1 or with negative SM at prostatectomy. Men with involved surgical margins and PSA <0.5 ng ml-1 appear to be at a lower risk of failure after SRT alone and may not derive as much benefit from the administration of ADT with SRT. These results are hypothesis-generating only, and further prospective data are required to see if ADT can safely be omitted in this select group of men.

Immunophenotypic and DNA genotypic analysis of T-cell and NK-cell subpopulations in patients with B-cell chronic lymphocytic leukaemia (B-CLL).

Absolute numbers and distributions of peripheral blood T-cells and NK cells were immunophenotypically determined in 21 patients with B-CLL and compared with those obtained from a series of 13 elderly normal controls with an age range of 60-87 years. For absolute CD3+, CD4+ and CD8+ T-cell, and CD16+ NK subpopulation numbers, there were no consistent differences between the normal and B-CLL groups although some individual patient variation was seen. Immunophenotypic analyses did however reveal that CD3+ T-cells in almost half (10/21) of the B-CLL patients were Ia+ (defined as > 20% positive cells), compared to 0/13 of the elderly control group (p < 0.001), and that the proportions of CD4+ and CD8+ cells expressing membrane CD45RO were significantly increased compared to the control group. Subdivision of the B-CLL cases into those with low (< 20%) and high (> 20%) proportions of CD3+ T-cells co-expressing Ia further showed that CD45RO expression by CD4+ fractions was particularly prominent in the Ia+ subgroup, and that the relative increase of CD4+CD45RO+ cells was primarily a consequence of decreased absolute numbers of CD4+CD45RA+ lymphocytes. This study also examined extracted DNA from enriched CD3+ T-cell fractions (obtained by immunomagnetic bead selection in 9 of the B-CLL cases) by PCR analysis with two primers for the T-cell gamma gene locus. With the V gamma C (consensus) primer, 8/9 cases were polyclonal and the remaining case was oligoclonal. For comparison, 7/9 CD3+ fractions were oligoclonal with the V gamma 9 primer with the other two cases being polyclonal. No monoclonal CD3+ components were found. It is suggested that the observed increased Ia expression by CD3+ cells and the predominance of CD4+ cells expressing membrane CD45RO in patients with B-CLL may be of potential relevance to understanding the pathogenesis and patterns of disease progression.

Patterns of membrane CD45 isoform expression by leukaemic blasts and normal mature myeloid cells.

The membrane expression of CD45R isoforms by leukaemic blasts from 92 cases of acute myeloid (AML) and 12 cases of lymphoblastic leukaemia was analysed by single and two-colour flow cytometry. Compared to normal mature granulocytes, which invariably expressed UCHL1 (CD45RO) but not 2H4 (CD45RA), leukaemic myeloid blasts showed 2H4+ UCHL1- and 2H4- ++UCHL1- composite patterns of expression with the first of these phenotypes being associated in particular with the most immature AML subtype (M0). Similar analyses of blast cell fractions from monocytic AML variants, revealed wide variation in CD45R expression in cases of M4, whereas M5 leukaemias were typically 2H4+ ++UCHL1+ with minor 2H4- UCHL1- components. As most normal mature monocytes were also 2H4+ ++UCHL1+, this suggested that monocytic myeloid differentiation was primarily associated with the "acquisition" of UCHL1 and the development of 2H4/UCHL1 co-expressing cells. The expression of membrane CD45RA by leukaemic blasts is an abnormal characteristic and may be related to the maintenance of an undifferentiated state by malignant myeloid precursors.

The MHC class I associated beta 2-microglobulin (beta 2m) light chain is expressed in a molar excess over HLA-ABC and CD1 on the membrane of leukaemic B cells but not leukaemic T cells: evidence for further beta 2m-associated molecules.

Beta 2-microglobulin (beta 2m) constitutes the common light chain of both the MHC-encoded HLA-ABC molecules and a group of structurally related glycoproteins recognized by antibodies of the first cluster of differentiation (CD1a, CD1b and CD1c). These CD1 antigens appear similar to murine T1 and Qa molecules in terms of structure and tissue distribution, although the question of inter-species homology is controversial. A further group of alloantigens expressed predominantly on T cells has been reported however, with immunogenetic characteristics more closely analogous to the murine T1/Qa system than the CD1 antigens, although their precise identity remains ill-defined. Having previously shown that malignant B cells may express membrane CD1c, we examined leukaemic B-cells corresponding to early lymphoblastic differentiation (null- and common acute lymphoblastic leukaemia) through to the terminal plasma cell stage for the expression of other non-HLA class I beta 2m-associated molecules. It was found that leukaemic B-cells at intermediate/late stages of differentiation, represented by non-Hodgkin's lymphoma (B-NHL) and 'hairy-cell' leukaemia (HCL), had significantly higher beta 2m:HLA-ABC ratios than did the cells from other types of B-cell malignancy. Although leukaemic B cells with a demonstrable non HLA-ABC-associated beta 2m component expressed detectable levels of CD1c, and insignificant levels of CD1a and CD1b, the antigen density was insufficient to account for the excess beta 2m. In vitro stimulation of leukaemic B cells by phorbol ester substantially increased the expression of HLA-ABC and CD1c, but also accentuated further the difference between the expression of these molecules and that of beta 2m. There was no detectable beta 2m other than that associated with HLA-ABC and CD1 on the surface of malignant T cells by contrast. Our findings strongly support the existence, at certain stages of leukaemic B-cell differentiation, of an additional beta 2m component(s) other than that associated with HLA-ABC and CD1.

Patterns of Membrane Antigen Expression by AML Blasts: Quantitation and Histogram Analysis.

Leukaemic myeloid blasts from non-monocytic (M1-M3, n = 36) and monocytic (M4 and M5, n = 21) AML cases were examined for the expression of 12 different membrane determinants by flow cytometry. Data analyses for each antigen included the determination of (a) the mean fluorescence intensity for the whole blast cell population, (b) the relative levels of membrane fluorescence for individual events (cells), and (c) a conventional assessment of the proportion of cells staining positively (i.e. exceeding a pre-defined level of fluorescence). Three main types of staining histogram were observed and, of these, the most commonly seen (348/432 and 176/252 of non-monocytic and monocytic AML histograms respectively) was characterised by an homogenous distribution of staining intensities which did not exceed two log decades of fluorescence (S-type). The second staining pattern was characterised by a continuous spectrum of fluorescence which exceeded two log decades of fluorescence (SE-type), and the third pattern showed evidence of two leukaemic populations with different levels of fluorescent staining (BI-type). With the exception of occasional AML cases which expressed CD7 or CD19 with low staining intensity, the expression of lymphoid-associated membrane CD3, CD10, and CD22 by AML blasts was insignificant. For comparison, analysing the histogram patterns of expression for the myeloid and non-lineage associated membrane determinants revealed that CD11c, CD13, CD14, and CD38 were mainly of S- or SE-type for the non-monocytic AML variants, with a minor but significant proportion of such cases expressing CD33 (7/36), CD34 (6/36) and HLA-Dr (6/36) with a BI-type staining pattern. Similarly, histogram patterns for CD13, CD33, CD34 and CD38 expression by the monocytic AML variants were predominantly of S- or SE- type, with minor proportions of cases expressing CD11c (7/21), CD14 (10/21), and HLA-Dr with BI-type staining. Comparisons between the mean fluorescence staining intensities for the whole blast cell population and conventional positive versus negative delineations for each antigen studied further suggested that semi-quantitative measurements of fluorescent staining were more informative and potentially of greater relevance to the study and diagnostic assessment of acute myeloid leukaemia subtypes.

Flow cytometric analysis of membrane CD11b, CD11c and CD14 expression in acute myeloid leukaemia: relationships with monocytic subtypes and the concept of relative antigen expression.

Blast cells from 26 cases of acute myeloid leukaemia (AML) were examined, by single and "two-colour" flow cytometry, for relationships between membrane CD11b (monoclonal antibody OKM1), CD11c (KB90) and CD14 (Leu-M3). Increased expression of all three determinants was associated with myelomonocytic leukaemias, with their relative diagnostic value in discriminating monocytic (M4 and M5) from non-monocytic (M1, M2 and M3) subtypes being CD14 greater than CD11c greater than CD11b. However, the results also indicated, because of the heterogenous expression of CD11c in particular, and to a lesser extent CD11b, that the patterns or histograms of fluorescent staining were potentially more informative than an empirical subdivision of blasts into positive and negative subpopulations. In addition, analysis of phenotypic correlations by simultaneous two-colour fluorescence showed that the expression of CD11b and CD11c determinants by leukaemic myeloid blasts was highly correlated, in contrast to the expression of CD14 and CD11c which were relatively independent. Consequently, CD11c+ myeloid blasts almost always coexpressed CD11b whereas CD14+ cases of AML often comprised CD14+ CD11c+ and CD14+ CD11c- subpopulations. It is concluded from these observations that CD11c immunophenotyping is a useful supplementary investigation, particularly in CD14- cases of myelomonocytic leukaemia. However, it is also apparent that the presence of membrane CD11c per se is not lineage-specific and that the level of expression is perhaps a more discriminatory factor.

Diagnostic application of monoclonal antibody KB90 (CD11c) in acute myeloid leukaemia.

The expression of membrane CD11c by leukaemic blast cells was examined (indirect immunorosetting) in 75 cases of acute leukaemia (myeloid, n = 60; lymphoid, n = 15) and evaluated as a potential marker for the diagnostic discrimination between monocytic (AMML-M4 and AMoL-M5) and non-monocytic (M1, M2 and M3) AML subtypes. Preliminary studies of normal bone marrow cells indicated that CD11c expression was not restricted to cells of monocytic lineage but was also present, with apparent lower density, on significant proportions of mature and immature granulocytes. Examination of acute myeloid leukaemia (AML) subtypes revealed that the non-monocytic leukaemias (n = 33) were CD11c-, defined as less than 30% positive cells, whereas all but one of the AMML-M4 (n = 13) and AMoL-M5 (n = 14) cases were CD11c+. All 15 cases of lymphoblastic leukaemia (ALL) showed less than 5% CD11c+ blasts. Membrane CD11c expression was also compared to the more widely used markers of monocytic differentiation; cytoplasmic alpha-naphthyl acetate esterase (ANAE) and membrane CD14 expression. This analysis showed that all 13 AMML-M4 leukaemias studied, including seven cases that were CD14- and eight that were ANAE-, were CD11c+. In addition, the AMoL-M5 cases (all of which were ANAE+) could be phenotypically subdivided into CD11c+ CD14+ (n = 9), CD11c+ CD14- (n = 4) and CD11c- CD14- (n = 1) subgroups. The study also confirmed that the discriminitive ability and sensitivity of the immunorosetting procedure for the detection of membrane CD11c compared favourably to immunofluorescent staining intensities as measured by flow cytometry.

Immunophenotypic and enzymatic studies do not support the concept of mixed monocytic-granulocytic differentiation in acute promyelocytic leukaemia (M3): a study of 44 cases.

Leukaemic promyelocytes from 30 cases of hypergranular and 14 cases of hypogranular acute promyelocytic leukaemia (M3) were analysed for the presence of monocyte-associated characteristics to determine whether there was any evidence of mixed (hybrid) granulocytic-monocytic differentiation. Cytochemically, a high proportion of hypergranular cases showed significant alpha-naphthyl acetate esterase (ANAE) staining and simultaneous chloroacetate esterase, and ANAE expression by single cells was commonly seen. These atypical staining patterns were, however, not a feature of hypogranular cases. Immunophenotypic studies revealed that most hypergranular M3 cases were HLA-DR- and that monocyte-associated membrane CD14 expression was low in all cases tested. In addition, serum lysozyme concentrations (20 cases) were generally within the normal range and thus inconsistent with monocytic involvement in the leukaemic process. The significance of atypical ANAE staining of leukaemic promyelocytes was further examined by analysing ANAE isoenzyme components (defined by isoelectric focusing) in 11 cases. The patterns obtained (G1 and G2) were identical to those found in normal granulocytes and did not show any evidence of monocyte-associated esterase isoenzyme expression. On the basis of these findings, it is considered that the differentiation process in acute promyelocytic leukaemia is relatively well conserved and that the atypical esterase cytochemistry of hypergranular promyelocytes does not reflect their mixed lineage nature but is simply a consequence of increased granulation.

Diagnostic differentiation of chronic B-cell malignancies using monoclonal antibody L161 (CD1c).

The expression of membrane CD1c, as defined by monoclonal antibody L161, was examined on malignant lymphoid cells from 191 cases of chronic lymphoproliferative disease and on eight 'normal' enriched tonsil B-cell extracts. Of 79 cases of chronic lymphocytic leukaemia (CLL) studied, 77 showed low (less than 20% positive cells) CD1c expression whereas 63/71 (89%) cases of B-PLL, HCL and B-NHL showed increased CD1c+ (but not CD1a or CD1b) components. In contrast, malignancies corresponding to terminal stages of B-cell differentiation (immunocytoma and myeloma) generally showed low CD1c expression as did lymphoid cells from 10 cases of post-thymic malignancy. Although there was some correlation between the expression of membrane CD1c and immunoglobulin (SIg) light chain densities (P less than 0.001), it is relevant in diagnostic terms that seven cases of B-NHL with low SIg staining intensities more typically associated with CLL were CD1c+. CD1c expression was not, however, correlated with the presence of CD23 or FMC7 determinants but did show a similar pattern of expression to that previously reported for beta-2 microglobulin. Determination of cellular CD1c by APAAP immunocytochemistry confirmed the presence of higher antigen densities in malignant B-cells at intermediate/late stages of differentiation and this interpretation was further supported by the finding that the majority of phenotypically mature tonsil B-cells were also CD1c+. The determination of CD1c expression by malignant B-cells may therefore be of particular value in the diagnostic differentiation of chronic lymphoproliferative disorders.

Membrane transferrin receptor (TfR) and nuclear proliferation-associated Ki-67 expression in hemopoietic malignancies.

The expression of membrane transferrin receptors (TfRs), as defined by monoclonal antibody OKT9, and the nuclear proliferation-associated antigen Ki-67 were examined in 159 cases of hematological malignancy. Of the "chronic" B and T cell leukemias studied (n = 85), 61% showed less than 5% OKT9-positive cells and only 7% of cases were TfR+ (defined as greater than 20% positive cells). For comparison, the acute leukemias (n = 62) showed higher (p less than 0.001) TfR expression with 39% TfR+ cases, although there was considerable variation within diagnostic subgroups. Nuclear Ki-67 expression was generally insignificant (less than 1%) in chronic leukemias (78 of 88 cases), although two of eight B cell-type prolymphocytic leukemia and four of four cases of plasma cell leukemia showed greater than 10% Ki-67+ components. In contrast, 47% (31 of 66) acute leukemias had greater than 10% Ki-67+ cells, although there appeared to be no relationship between Ki-67 expression and leukemic type. Combined assessments of TfR and Ki-67 expression revealed a Ki-67- TfR- phenotype in 82% of chronic leukemias, compared with 28% of acute type, and a Ki-67+ TfR+ pattern was found in 27% of acute proliferations but not in any case of chronic leukemia. The determination of membrane TfR expression appears to have little value in the diagnostic differentiation of leukemias, whereas Ki-67 is considered to be a useful supplementary investigation in defining high grade tumors, in the recognition of prognostically poor cases of otherwise well defined low grade malignancy, and of potential value in resolving discrepancies between morphological and immunophenotypic features in leukemias of "intermediate" type.

Diagnostic and prognostic factors in acute monocytic leukaemia: an analysis of 51 cases.

Diagnostic features (cytochemistry, immunophenotyping and serum biochemistry) were examined in 51 cases of acute monocytic leukaemia (AMoL). Peroxidase, Sudan black B and alpha naphthyl acetate esterase (ANAE) cytochemical reactions were unrelated to morphological (FAB groups M5a and M5b) or immunological subtype. ANAE cytochemistry, however, indicated that AMoL cases could be subdivided into those with typical (M-type) reactions and those with insignificant staining or monocytic ANAE isoenzymes (defined by IEF). All cases were phenotypically CD13/CD33 positive and, with one exception, had greater than 30% HLA-DR positive cells. Membrane CD14 expression was insignificant or variable in 33% of M5a cases in contrast to 23/24 M5b cases which showed high proportions of CD14-staining cells with at least two monoclonal antibodies. Serum lysozyme, LDH and beta-2 microglobulin (beta 2m) were increased in 88%, 68% and 81% of cases respectively but, with the exception of statistically higher lysozyme levels in CD14+ cases, were unrelated to the morphological, cytochemical or immunological diagnostic subgroups. Clinical and diagnostic features were also examined as possible prognostic indicators. The morphological, cytochemical and immunological subgroups of AMoL were not found to be of prognostic relevance but age (P = 0.004), renal failure (P = 0.005) and serum beta 2m levels (P = 0.002) were related to patient survival. Moreover, renal failure and serum beta 2m remained significant (P = 0.012 respectively) when age was taken into account and were shown to be independent prognostic variables.